----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
      name:  <unnamed>
       log:  C:\Users\marko\Dropbox\07_LatinAmerica_Corruption_Paper\LAPOP\replication\appendix-log.log
  log type:  text
 opened on:  12 May 2020, 09:30:44

. 
. ***************************************************************
. * Klasnja, Lupu, Tucker: 
. * "When Do Voters Sanction Corrupt Politicians?"
. * This do file replicates the results show in the appendix. 
. ***************************************************************
. 
. * for further details, see the readme.txt file in the replication archive
. * user-written command -coefplot- needed; type 'ssc install coefplot'
. * user-written command -frmttable- needed; type 'ssc install frmttable'
. * user-written command -estout- needed; type 'ssc install estout'
. * user-written plot scheme -plotplain- needed; type 'net install gr0070'
. 
. *** Table A1: Summary statistics
. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. mat n = J(1,4,.)

. mat sum = J(9,4,.)

. bys uniq_id: egen count = seq(), from(1)

. keep if count == 1
(4,667 observations deleted)

. qui sum uniq_id

. mat n1 = r(N)

. tab country, matcell(n2)

    Country |      Freq.     Percent        Cum.
------------+-----------------------------------
  Argentina |      1,528       32.74       32.74
      Chile |      1,625       34.82       67.56
    Uruguay |      1,514       32.44      100.00
------------+-----------------------------------
      Total |      4,667      100.00

. mat n2 = n2'

. mat n = n1,n2

. gen corperchi1 = (corperc > 3) if corperc ~= .
(177 missing values generated)

. preserve

. collapse r_female r_age loed2 lowealth unemp bribejust corperchi1 corperchi anybribe [pw=weight1500]

. mkmat r_female-anybribe, mat(s1)

. mat s1 = s1'

. restore

. collapse r_female r_age loed2 lowealth unemp bribejust corperchi1 corperchi anybribe [pw=weight1500], by(country)

. mkmat r_female-anybribe, mat(s2)

. mat s2 = s2'

. mat sum = s1,s2

. mat all = sum \ n

. frmttable, statmat(all) substat(0) sdec(2 \ 2 \ 2 \ 2 \ 2 \ 2 \ 2 \ 2 \ 2 \ 0) ///
>         rtitles("Female" \ "Age" \ "Less than high school" \ "Wealth (in bottom 3 quintiles)" \ ///
>                 "Unemployment" \ "Bribes justifiable" \ "High corruption perception" \ ///
>                 "Very high corruption perception" \ ///
>                 "Bribe paid" \ "Observations") ///
>         ctitles("","All","Argentina","Chile","Uruguay")

                                                            ---------------------------------------------------------------------
                                                                                                All   Argentina  Chile  Uruguay 
                                                            ---------------------------------------------------------------------
                                                             Female                            0.51     0.50     0.51    0.52   
                                                             Age                               43.53    41.90    42.18   46.51  
                                                             Less than high school             0.79     0.78     0.78    0.81   
                                                             Wealth (in bottom 3 quintiles)    0.60     0.59     0.61    0.59   
                                                             Unemployment                      0.13     0.09     0.16    0.14   
                                                             Bribes justifiable                0.12     0.10     0.17    0.10   
                                                             High corruption perception        0.60     0.66     0.72    0.40   
                                                             Very high corruption perception   0.22     0.26     0.29    0.10   
                                                             Bribe paid                        0.10     0.16     0.07    0.06   
                                                             Observations                      4,667    1,528    1,625   1,514  
                                                            ---------------------------------------------------------------------


. frmttable, clear

. 
. 
. *** Figure A1: Diagnostic check for profile order effects
. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. * incumbent
. reg vote i.corrupttreat badeconomy i.copartisan i.copsource2 i.female i.country if arias == 0 [pw=weight1500], cl(uniq_id)
(sum of wgt is 4,177.53590518236)

Linear regression                               Number of obs     =      4,334
                                                F(10, 4333)       =      50.83
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1136
                                                Root MSE          =     .42824

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3496083   .0189552   -18.44   0.000    -.3867701   -.3124465
         Bribes common  |  -.3492641   .0191849   -18.21   0.000    -.3868763   -.3116518
       Bribes but jobs  |  -.2652565   .0205029   -12.94   0.000    -.3054528   -.2250603
                        |
             badeconomy |  -.0675986   .0130197    -5.19   0.000    -.0931238   -.0420735
                        |
             copartisan |
 Co-partisan candidate  |   .0939774   .0263758     3.56   0.000     .0422674    .1456874
                        |
             copsource2 |
Out-partisan newspaper  |   .0149276   .0145213     1.03   0.304    -.0135416    .0433967
 Co-partisan newspaper  |    .032405   .0185289     1.75   0.080    -.0039212    .0687312
                        |
                 female |
                Female  |  -.0452229   .0130153    -3.47   0.001    -.0707395   -.0197062
                        |
                country |
                 Chile  |     .00807   .0158546     0.51   0.611    -.0230131    .0391531
               Uruguay  |   .0368595   .0162195     2.27   0.023      .005061    .0686581
                        |
                  _cons |   .5552604    .022401    24.79   0.000     .5113431    .5991778
-----------------------------------------------------------------------------------------

. estimates store r1

. * challenger
. reg vote i.corrupttreat badeconomy i.copartisan i.copsource2 i.female i.country if arias == 1 [pw=weight1500], cl(uniq_id)
(sum of wgt is 4,177.53590518236)

Linear regression                               Number of obs     =      4,334
                                                F(10, 4333)       =      47.81
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1118
                                                Root MSE          =     .42643

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3572634   .0192263   -18.58   0.000    -.3949567     -.31957
         Bribes common  |  -.3758801   .0190194   -19.76   0.000    -.4131678   -.3385924
       Bribes but jobs  |  -.2642395   .0203878   -12.96   0.000    -.3042101    -.224269
                        |
             badeconomy |   .0290966   .0129852     2.24   0.025     .0036391    .0545542
                        |
             copartisan |
 Co-partisan candidate  |   .0566595   .0286767     1.98   0.048     .0004385    .1128805
                        |
             copsource2 |
Out-partisan newspaper  |  -.0190041   .0145751    -1.30   0.192    -.0475788    .0095706
 Co-partisan newspaper  |   .0234093   .0187028     1.25   0.211    -.0132579    .0600764
                        |
                 female |
                Female  |     .00725   .0129726     0.56   0.576     -.018183     .032683
                        |
                country |
                 Chile  |  -.0255711   .0160149    -1.60   0.110    -.0569685    .0058264
               Uruguay  |  -.0326752   .0162703    -2.01   0.045    -.0645732   -.0007772
                        |
                  _cons |   .5396251   .0222762    24.22   0.000     .4959524    .5832978
-----------------------------------------------------------------------------------------

. estimates store r2

. coefplot (r1, label(Incumbent) msymbol(T) msize(medsmall) mcolor(black%80) offset(0.2)) ///
>         (r2, label(Challenger) msymbol(S) msize(medsmall) mcolor(gray%80) offset(-0.2)), ///
>         scheme(plotplain) xline(0, lcolor(black) lpattern(solid) lwidth(vthin)) baselevels ///
>         graphregion(color(white)) drop(_cons 1.country 2.country 3.country) legend(ring(0) position(1)) ///
>                 xtitle("Effect on Pr(Voting for candidate)") coeflabels(, labsize(small)) grid(glcolor(gs2)) ///
>                 headings(1.corrupttreat = "{bf:Corruption}" ///
>                         0.badeconomy = "{bf:Economy}" ///
>                         0.female = "{bf: Gender}" ///
>                         0.copartisan = "{bf:Partisanship}" ///
>                         0.copsource2 = "{bf:Information source}")

. 
. 
. *** Table A2: Results with economy and gender interactions
. eststo clear

. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. eststo: reg vote i.corrupttreat i.copartisan i.badeconomy##inc i.copsource2 i.female i.country ///
>         [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(12, 4333)       =      82.21
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1103
                                                Root MSE          =      .4277

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3533974    .013401   -26.37   0.000    -.3796702   -.3271247
         Bribes common  |  -.3631275   .0134214   -27.06   0.000    -.3894404   -.3368146
       Bribes but jobs  |   -.264988   .0145094   -18.26   0.000    -.2934338   -.2365422
                        |
             copartisan |
 Co-partisan candidate  |   .0775461   .0188731     4.11   0.000     .0405451     .114547
                        |
             badeconomy |
      Economy worsened  |   .0291838   .0129845     2.25   0.025     .0037275      .05464
                  1.inc |   .0528361   .0153538     3.44   0.001     .0227349    .0829373
                        |
         badeconomy#inc |
    Economy worsened#1  |  -.0968282    .021476    -4.51   0.000    -.1389322   -.0547242
                        |
             copsource2 |
Out-partisan newspaper  |  -.0015192   .0103112    -0.15   0.883    -.0217345    .0186961
 Co-partisan newspaper  |   .0283579   .0130783     2.17   0.030     .0027178    .0539981
                        |
                 female |
                Female  |  -.0191172   .0092638    -2.06   0.039    -.0372791   -.0009553
                        |
                country |
                 Chile  |  -.0083276   .0090087    -0.92   0.355    -.0259894    .0093341
               Uruguay  |   .0021432   .0091276     0.23   0.814    -.0157516    .0200379
                        |
                  _cons |   .5207411   .0163245    31.90   0.000     .4887368    .5527454
-----------------------------------------------------------------------------------------
(est1 stored)

. eststo: reg vote (i.corrupttreat i.copartisan inc i.copsource2 i.female i.country)##i.badeconomy ///
>         [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(21, 4333)       =      48.25
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1118
                                                Root MSE          =     .42755

                                                        (Std. Err. adjusted for 4,334 clusters in uniq_id)
----------------------------------------------------------------------------------------------------------
                                         |               Robust
                                    vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-----------------------------------------+----------------------------------------------------------------
                            corrupttreat |
                                 Bribes  |  -.3589633   .0189711   -18.92   0.000    -.3961564   -.3217703
                          Bribes common  |  -.3671957   .0189755   -19.35   0.000    -.4043974   -.3299941
                        Bribes but jobs  |  -.2737291   .0205034   -13.35   0.000    -.3139263   -.2335319
                                         |
                              copartisan |
                  Co-partisan candidate  |   .0918047   .0277013     3.31   0.001      .037496    .1461134
                                   1.inc |   .0524284   .0153365     3.42   0.001     .0223611    .0824957
                                         |
                              copsource2 |
                 Out-partisan newspaper  |  -.0161071   .0150906    -1.07   0.286    -.0456925    .0134782
                  Co-partisan newspaper  |  -.0159101   .0180537    -0.88   0.378    -.0513046    .0194843
                                         |
                                  female |
                                 Female  |  -.0269787    .013159    -2.05   0.040    -.0527772   -.0011802
                                         |
                                 country |
                                  Chile  |  -.0129571    .012835    -1.01   0.313    -.0381204    .0122061
                                Uruguay  |  -.0061061   .0128732    -0.47   0.635    -.0313442    .0191319
                                         |
                              badeconomy |
                       Economy worsened  |  -.0290848   .0297708    -0.98   0.329    -.0874509    .0292812
                                         |
                 corrupttreat#badeconomy |
                Bribes#Economy worsened  |   .0115646    .026811     0.43   0.666    -.0409987    .0641278
         Bribes common#Economy worsened  |   .0083406   .0268512     0.31   0.756    -.0443015    .0609827
       Bribes but jobs#Economy worsened  |   .0162375   .0290267     0.56   0.576    -.0406695    .0731446
                                         |
                   copartisan#badeconomy |
 Co-partisan candidate#Economy worsened  |  -.0293668   .0377656    -0.78   0.437    -.1034067    .0446731
                                         |
                          inc#badeconomy |
                     1#Economy worsened  |  -.0958084   .0214702    -4.46   0.000    -.1379009   -.0537159
                                         |
                   copsource2#badeconomy |
Out-partisan newspaper#Economy worsened  |   .0293803     .02063     1.42   0.154    -.0110649    .0698256
 Co-partisan newspaper#Economy worsened  |   .0920597   .0261844     3.52   0.000     .0407249    .1433945
                                         |
                       female#badeconomy |
                Female#Economy worsened  |   .0162624   .0185206     0.88   0.380    -.0200475    .0525722
                                         |
                      country#badeconomy |
                 Chile#Economy worsened  |   .0115565   .0180622     0.64   0.522    -.0238546    .0469675
               Uruguay#Economy worsened  |   .0201261   .0182773     1.10   0.271    -.0157066    .0559589
                                         |
                                   _cons |   .5487912   .0211094    26.00   0.000     .5074059    .5901765
----------------------------------------------------------------------------------------------------------
(est2 stored)

. eststo: reg vote (i.corrupttreat i.copartisan i.badeconomy##inc i.copsource2 i.country)##i.female ///
>         [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(23, 4333)       =      43.99
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1125
                                                Root MSE          =     .42743

                                              (Std. Err. adjusted for 4,334 clusters in uniq_id)
------------------------------------------------------------------------------------------------
                               |               Robust
                          vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------------------------+----------------------------------------------------------------
                  corrupttreat |
                       Bribes  |  -.3359663   .0188219   -17.85   0.000    -.3728668   -.2990658
                Bribes common  |  -.3462448   .0186119   -18.60   0.000    -.3827337    -.309756
              Bribes but jobs  |  -.2448667    .020417   -11.99   0.000    -.2848945    -.204839
                               |
                    copartisan |
        Co-partisan candidate  |   .0915421   .0269122     3.40   0.001     .0387804    .1443039
                               |
                    badeconomy |
             Economy worsened  |   .0309392   .0183794     1.68   0.092    -.0050938    .0669722
                         1.inc |   .0881586   .0204635     4.31   0.000     .0480396    .1282776
                               |
                badeconomy#inc |
           Economy worsened#1  |  -.1165157   .0285579    -4.08   0.000    -.1725038   -.0605275
                               |
                    copsource2 |
       Out-partisan newspaper  |  -.0153571   .0145199    -1.06   0.290    -.0438236    .0131093
        Co-partisan newspaper  |    .013309   .0189397     0.70   0.482    -.0238226    .0504406
                               |
                       country |
                        Chile  |  -.0184646   .0145175    -1.27   0.203    -.0469264    .0099971
                      Uruguay  |   .0170822   .0148778     1.15   0.251    -.0120859    .0462504
                               |
                        female |
                       Female  |   .0246413   .0331058     0.74   0.457     -.040263    .0895456
                               |
           corrupttreat#female |
                Bribes#Female  |  -.0375188   .0267138    -1.40   0.160    -.0898914    .0148539
         Bribes common#Female  |  -.0358077   .0267178    -1.34   0.180    -.0881881    .0165728
       Bribes but jobs#Female  |  -.0413071   .0289616    -1.43   0.154    -.0980867    .0154725
                               |
             copartisan#female |
 Co-partisan candidate#Female  |  -.0286174    .038207    -0.75   0.454    -.1035227    .0462879
                               |
             badeconomy#female |
      Economy worsened#Female  |  -.0033733   .0259784    -0.13   0.897    -.0543043    .0475576
                               |
                    inc#female |
                     1#Female  |   -.072239   .0259298    -2.79   0.005    -.1230747   -.0214032
                               |
         badeconomy#inc#female |
    Economy worsened#1#Female  |   .0405672   .0364525     1.11   0.266    -.0308984    .1120327
                               |
             copsource2#female |
Out-partisan newspaper#Female  |   .0277983   .0205194     1.35   0.176    -.0124302    .0680267
 Co-partisan newspaper#Female  |   .0292942   .0265094     1.11   0.269    -.0226778    .0812663
                               |
                country#female |
                 Chile#Female  |    .018784   .0225786     0.83   0.405    -.0254817    .0630497
               Uruguay#Female  |  -.0305509   .0233072    -1.31   0.190    -.0762449    .0151431
                               |
                         _cons |   .5003542   .0223395    22.40   0.000     .4565573    .5441511
------------------------------------------------------------------------------------------------
(est3 stored)

. esttab, label b(2) se(2) nonumbers nobaselevels nolegend star(* 0.10 ** 0.05 *** 0.01) ///
>         mtitles("Main model" "Interactions w/ economy" "Interactions w/ female") drop(2.country 3.country) ///
>         wide

-----------------------------------------------------------------------------------------------------------
                       Main model                 Interactio~y                 Interactio~e                
-----------------------------------------------------------------------------------------------------------
Bribes                      -0.35***       (0.01)        -0.36***       (0.02)        -0.34***       (0.02)
Bribes common               -0.36***       (0.01)        -0.37***       (0.02)        -0.35***       (0.02)
Bribes but jobs             -0.26***       (0.01)        -0.27***       (0.02)        -0.24***       (0.02)
Co-partisan candid~e         0.08***       (0.02)         0.09***       (0.03)         0.09***       (0.03)
Economy worsened             0.03**        (0.01)        -0.03          (0.03)         0.03*         (0.02)
Experiment: Incumb~1         0.05***       (0.02)         0.05***       (0.02)         0.09***       (0.02)
Economy worsened #~I        -0.10***       (0.02)        -0.10***       (0.02)        -0.12***       (0.03)
Out-partisan newsp~r        -0.00          (0.01)        -0.02          (0.02)        -0.02          (0.01)
Co-partisan newspa~r         0.03**        (0.01)        -0.02          (0.02)         0.01          (0.02)
Female                      -0.02**        (0.01)        -0.03**        (0.01)         0.02          (0.03)
Bribes # Economy w~d                                      0.01          (0.03)                             
Bribes common # Ec~d                                      0.01          (0.03)                             
Bribes but jobs # ~n                                      0.02          (0.03)                             
Co-partisan candid~                                      -0.03          (0.04)                             
Out-partisan newsp~y                                      0.03          (0.02)                             
Co-partisan newspa~                                       0.09***       (0.03)                             
Female # Economy w~d                                      0.02          (0.02)        -0.00          (0.03)
Chile # Economy wo~d                                      0.01          (0.02)                             
Uruguay # Economy ~d                                      0.02          (0.02)                             
Bribes # Female                                                                       -0.04          (0.03)
Bribes common # Fe~e                                                                  -0.04          (0.03)
Bribes but jobs # ~e                                                                  -0.04          (0.03)
Co-partisan candid~e                                                                  -0.03          (0.04)
Experiment: Incum~1                                                                   -0.07***       (0.03)
Economy worsened #~I                                                                   0.04          (0.04)
Out-partisan newsp~e                                                                   0.03          (0.02)
Co-partisan newspa~e                                                                   0.03          (0.03)
Chile # Female                                                                         0.02          (0.02)
Uruguay # Female                                                                      -0.03          (0.02)
Constant                     0.52***       (0.02)         0.55***       (0.02)         0.50***       (0.02)
-----------------------------------------------------------------------------------------------------------
Observations                 8668                         8668                         8668                
-----------------------------------------------------------------------------------------------------------
Standard errors in parentheses

.         
. 
. *** Table A3: Randomization checks
. eststo clear

. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. eststo: reg r_female i.corrupttreat i.copartisan i.badeconomy##i.inc i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,994.34440624714)

Linear regression                               Number of obs     =      9,328
                                                F(12, 4663)       =       1.73
                                                Prob > F          =     0.0540
                                                R-squared         =     0.0027
                                                Root MSE          =     .49955

                                       (Std. Err. adjusted for 4,664 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
               r_female |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |   .0124684   .0147009     0.85   0.396    -.0163523    .0412891
         Bribes common  |   .0043982   .0148732     0.30   0.767    -.0247603    .0335566
       Bribes but jobs  |   .0214629   .0149901     1.43   0.152    -.0079248    .0508506
                        |
             copartisan |
 Co-partisan candidate  |  -.0349665   .0240857    -1.45   0.147    -.0821858    .0122527
                        |
             badeconomy |
      Economy worsened  |  -.0119312   .0146266    -0.82   0.415    -.0406062    .0167438
                  1.inc |   .0004176   .0007644     0.55   0.585    -.0010811    .0019162
                        |
         badeconomy#inc |
    Economy worsened#1  |   .0012138   .0010306     1.18   0.239    -.0008067    .0032343
                        |
             copsource2 |
Out-partisan newspaper  |    .018161   .0118106     1.54   0.124    -.0049935    .0413154
 Co-partisan newspaper  |  -.0199262   .0150915    -1.32   0.187    -.0495127    .0096602
                        |
                 female |
                Female  |   .0272101   .0102514     2.65   0.008     .0071125    .0473076
                        |
                country |
                 Chile  |     .00084    .017816     0.05   0.962    -.0340879    .0357679
               Uruguay  |   .0245475   .0182607     1.34   0.179    -.0112521    .0603471
                        |
                  _cons |    .483354   .0196507    24.60   0.000     .4448292    .5218788
-----------------------------------------------------------------------------------------
(est1 stored)

. eststo: reg r_age i.corrupttreat i.copartisan i.badeconomy##i.inc i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,992.49825239182)

Linear regression                               Number of obs     =      9,326
                                                F(12, 4662)       =       8.62
                                                Prob > F          =     0.0000
                                                R-squared         =     0.0202
                                                Root MSE          =     17.153

                                       (Std. Err. adjusted for 4,663 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                  r_age |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |   .5427089   .5088529     1.07   0.286    -.4548835    1.540301
         Bribes common  |   .1975995   .5089603     0.39   0.698    -.8002034    1.195402
       Bribes but jobs  |   .4618217   .5166563     0.89   0.371     -.551069    1.474712
                        |
             copartisan |
 Co-partisan candidate  |   4.814409   .8654886     5.56   0.000     3.117642    6.511176
                        |
             badeconomy |
      Economy worsened  |   .6619403   .5039004     1.31   0.189    -.3259428    1.649823
                  1.inc |  -.0483147   .0349149    -1.38   0.166    -.1167644     .020135
                        |
         badeconomy#inc |
    Economy worsened#1  |  -.0274653   .0471117    -0.58   0.560    -.1198266    .0648959
                        |
             copsource2 |
Out-partisan newspaper  |   .1212594    .405782     0.30   0.765    -.6742653    .9167841
 Co-partisan newspaper  |   .6652416   .5198987     1.28   0.201    -.3540058    1.684489
                        |
                 female |
                Female  |  -.2280724   .3559298    -0.64   0.522    -.9258632    .4697184
                        |
                country |
                 Chile  |    .295551   .6014175     0.49   0.623    -.8835117    1.474614
               Uruguay  |   4.126016   .6427294     6.42   0.000     2.865962    5.386069
                        |
                  _cons |   41.06698    .671234    61.18   0.000     39.75104    42.38291
-----------------------------------------------------------------------------------------
(est2 stored)

. eststo: reg ed i.corrupttreat i.copartisan i.badeconomy##i.inc i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 7,757.64592850208)

Linear regression                               Number of obs     =      8,038
                                                F(12, 4018)       =       1.58
                                                Prob > F          =     0.0898
                                                R-squared         =     0.0031
                                                Root MSE          =     1.8743

                                       (Std. Err. adjusted for 4,019 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   educ |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |   .0199485   .0594698     0.34   0.737    -.0966452    .1365422
         Bribes common  |  -.0590586   .0602178    -0.98   0.327    -.1771189    .0590018
       Bribes but jobs  |   .0211377   .0596289     0.35   0.723     -.095768    .1380434
                        |
             copartisan |
 Co-partisan candidate  |   .0647997   .0977595     0.66   0.507    -.1268633    .2564626
                        |
             badeconomy |
      Economy worsened  |  -.0902528   .0591858    -1.52   0.127    -.2062898    .0257842
                  1.inc |  -.0023861    .002709    -0.88   0.378    -.0076971     .002925
                        |
         badeconomy#inc |
    Economy worsened#1  |  -.0002013   .0037014    -0.05   0.957     -.007458    .0070554
                        |
             copsource2 |
Out-partisan newspaper  |  -.1299941   .0474237    -2.74   0.006    -.2229709   -.0370173
 Co-partisan newspaper  |  -.0768067   .0607921    -1.26   0.207     -.195993    .0423795
                        |
                 female |
                Female  |   .0298742   .0418676     0.71   0.476    -.0522096     .111958
                        |
                country |
                 Chile  |    -.14395   .0735571    -1.96   0.050    -.2881627    .0002627
               Uruguay  |  -.1330341   .0719568    -1.85   0.065    -.2741094    .0080413
                        |
                  _cons |   2.779067   .0790762    35.14   0.000     2.624033      2.9341
-----------------------------------------------------------------------------------------
(est3 stored)

. esttab, label b(2) se(2) nonumbers nobaselevels nolegend star(* 0.10 ** 0.05 *** 0.01) ///
>         mtitles("Gender" "Age" "Education") drop(2.country 3.country _cons)

--------------------------------------------------------------------
                           Gender             Age       Education   
--------------------------------------------------------------------
Bribes                       0.01            0.54            0.02   
                           (0.01)          (0.51)          (0.06)   

Bribes common                0.00            0.20           -0.06   
                           (0.01)          (0.51)          (0.06)   

Bribes but jobs              0.02            0.46            0.02   
                           (0.01)          (0.52)          (0.06)   

Co-partisan candid~e        -0.03            4.81***         0.06   
                           (0.02)          (0.87)          (0.10)   

Economy worsened            -0.01            0.66           -0.09   
                           (0.01)          (0.50)          (0.06)   

Experiment: Incumb~1         0.00           -0.05           -0.00   
                           (0.00)          (0.03)          (0.00)   

Economy worsened #~I         0.00           -0.03           -0.00   
                           (0.00)          (0.05)          (0.00)   

Out-partisan newsp~r         0.02            0.12           -0.13***
                           (0.01)          (0.41)          (0.05)   

Co-partisan newspa~r        -0.02            0.67           -0.08   
                           (0.02)          (0.52)          (0.06)   

Female                       0.03***        -0.23            0.03   
                           (0.01)          (0.36)          (0.04)   
--------------------------------------------------------------------
Observations                 9328            9326            8038   
--------------------------------------------------------------------
Standard errors in parentheses

.         
.         
. *** Table A4: Coefficient Estimates for Figures 1-3 in the main text
. eststo clear

. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. eststo: reg vote i.corrupttreat i.copartisan i.badeconomy##i.inc i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(12, 4333)       =      82.21
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1103
                                                Root MSE          =      .4277

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3533974    .013401   -26.37   0.000    -.3796702   -.3271247
         Bribes common  |  -.3631275   .0134214   -27.06   0.000    -.3894404   -.3368146
       Bribes but jobs  |   -.264988   .0145094   -18.26   0.000    -.2934338   -.2365422
                        |
             copartisan |
 Co-partisan candidate  |   .0775461   .0188731     4.11   0.000     .0405451     .114547
                        |
             badeconomy |
      Economy worsened  |   .0291838   .0129845     2.25   0.025     .0037275      .05464
                  1.inc |   .0528361   .0153538     3.44   0.001     .0227349    .0829373
                        |
         badeconomy#inc |
    Economy worsened#1  |  -.0968282    .021476    -4.51   0.000    -.1389322   -.0547242
                        |
             copsource2 |
Out-partisan newspaper  |  -.0015192   .0103112    -0.15   0.883    -.0217345    .0186961
 Co-partisan newspaper  |   .0283579   .0130783     2.17   0.030     .0027178    .0539981
                        |
                 female |
                Female  |  -.0191172   .0092638    -2.06   0.039    -.0372791   -.0009553
                        |
                country |
                 Chile  |  -.0083276   .0090087    -0.92   0.355    -.0259894    .0093341
               Uruguay  |   .0021432   .0091276     0.23   0.814    -.0157516    .0200379
                        |
                  _cons |   .5207411   .0163245    31.90   0.000     .4887368    .5527454
-----------------------------------------------------------------------------------------
(est1 stored)

. eststo: reg vote i.corrupttreat##i.copartisan i.badeconomy##i.inc i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(15, 4333)       =      66.34
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1110
                                                Root MSE          =     .42759

                                                      (Std. Err. adjusted for 4,334 clusters in uniq_id)
--------------------------------------------------------------------------------------------------------
                                       |               Robust
                                  vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
---------------------------------------+----------------------------------------------------------------
                          corrupttreat |
                               Bribes  |  -.3462265   .0139115   -24.89   0.000    -.3735001   -.3189529
                        Bribes common  |     -.3582   .0139468   -25.68   0.000    -.3855428   -.3308572
                      Bribes but jobs  |  -.2561961   .0149986   -17.08   0.000    -.2856011   -.2267911
                                       |
                            copartisan |
                Co-partisan candidate  |   .1550649   .0416301     3.72   0.000     .0734486    .2366811
                                       |
               corrupttreat#copartisan |
         Bribes#Co-partisan candidate  |  -.1085931   .0552178    -1.97   0.049    -.2168482    -.000338
  Bribes common#Co-partisan candidate  |  -.0727644   .0548904    -1.33   0.185    -.1803776    .0348489
Bribes but jobs#Co-partisan candidate  |  -.1314282   .0597686    -2.20   0.028    -.2486053   -.0142512
                                       |
                            badeconomy |
                     Economy worsened  |   .0289751   .0129864     2.23   0.026     .0035152     .054435
                                 1.inc |   .0526438   .0153534     3.43   0.001     .0225432    .0827443
                                       |
                        badeconomy#inc |
                   Economy worsened#1  |  -.0966118   .0214637    -4.50   0.000    -.1386916   -.0545319
                                       |
                            copsource2 |
               Out-partisan newspaper  |  -.0006011   .0103238    -0.06   0.954    -.0208411    .0196389
                Co-partisan newspaper  |   .0289511   .0130758     2.21   0.027     .0033158    .0545863
                                       |
                                female |
                               Female  |  -.0190311   .0092637    -2.05   0.040    -.0371927   -.0008694
                                       |
                               country |
                                Chile  |  -.0084767   .0090123    -0.94   0.347    -.0261455     .009192
                              Uruguay  |   .0021014   .0091422     0.23   0.818    -.0158219    .0200247
                                       |
                                 _cons |   .5150949   .0165245    31.17   0.000     .4826985    .5474913
--------------------------------------------------------------------------------------------------------
(est2 stored)

. eststo: reg vote i.corrupttreat##i.bribejust i.copartisan i.badeconomy##i.inc i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,256.80455803871)

Linear regression                               Number of obs     =      8,566
                                                F(16, 4282)       =      66.33
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1162
                                                Root MSE          =      .4267

                                               (Std. Err. adjusted for 4,283 clusters in uniq_id)
-------------------------------------------------------------------------------------------------
                                |               Robust
                           vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
--------------------------------+----------------------------------------------------------------
                   corrupttreat |
                        Bribes  |  -.3666128   .0142513   -25.72   0.000    -.3945528   -.3386728
                 Bribes common  |   -.367043   .0144461   -25.41   0.000    -.3953648   -.3387212
               Bribes but jobs  |   -.284107   .0154706   -18.36   0.000    -.3144375   -.2537766
                                |
                      bribejust |
                High tolerance  |   .0160294   .0299675     0.53   0.593    -.0427224    .0747812
                                |
         corrupttreat#bribejust |
         Bribes#High tolerance  |   .0846854   .0431845     1.96   0.050     .0000214    .1693495
  Bribes common#High tolerance  |   .0033493   .0409998     0.08   0.935    -.0770315    .0837301
Bribes but jobs#High tolerance  |   .1241313   .0452667     2.74   0.006      .035385    .2128776
                                |
                     copartisan |
         Co-partisan candidate  |   .0793506   .0188633     4.21   0.000     .0423687    .1163325
                                |
                     badeconomy |
              Economy worsened  |   .0300843   .0130516     2.31   0.021     .0044965    .0556722
                          1.inc |   .0524429   .0154359     3.40   0.001     .0221804    .0827054
                                |
                 badeconomy#inc |
            Economy worsened#1  |  -.0982962   .0215829    -4.55   0.000    -.1406098   -.0559827
                                |
                     copsource2 |
        Out-partisan newspaper  |  -.0043259   .0103424    -0.42   0.676    -.0246023    .0159505
         Co-partisan newspaper  |   .0249166   .0131384     1.90   0.058    -.0008415    .0506748
                                |
                         female |
                        Female  |  -.0192951   .0092958    -2.08   0.038    -.0375198   -.0010705
                                |
                        country |
                         Chile  |   -.010226   .0090496    -1.13   0.259    -.0279678    .0075159
                       Uruguay  |   .0011985    .009163     0.13   0.896    -.0167658    .0191628
                                |
                          _cons |   .5249562   .0169742    30.93   0.000      .491678    .5582344
-------------------------------------------------------------------------------------------------
(est3 stored)

. eststo: reg vote i.corrupttreat##i.loed2 i.copartisan i.badeconomy##i.inc i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(16, 4333)       =      65.13
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1144
                                                Root MSE          =      .4268

                                              (Std. Err. adjusted for 4,334 clusters in uniq_id)
------------------------------------------------------------------------------------------------
                               |               Robust
                          vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------------------------+----------------------------------------------------------------
                  corrupttreat |
                       Bribes  |  -.4594453   .0278717   -16.48   0.000    -.5140881   -.4048026
                Bribes common  |  -.4881154   .0283177   -17.24   0.000    -.5436327   -.4325982
              Bribes but jobs  |  -.4203397   .0304641   -13.80   0.000     -.480065   -.3606144
                               |
                         loed2 |
                Low education  |  -.1244136   .0243211    -5.12   0.000    -.1720954   -.0767318
                               |
            corrupttreat#loed2 |
         Bribes#Low education  |    .133302   .0317575     4.20   0.000     .0710411     .195563
  Bribes common#Low education  |   .1572822   .0321129     4.90   0.000     .0943244    .2202399
Bribes but jobs#Low education  |    .194445   .0345698     5.62   0.000     .1266705    .2622195
                               |
                    copartisan |
        Co-partisan candidate  |   .0794764   .0188686     4.21   0.000     .0424844    .1164685
                               |
                    badeconomy |
             Economy worsened  |   .0283106   .0129579     2.18   0.029     .0029064    .0537148
                         1.inc |   .0525729    .015327     3.43   0.001     .0225242    .0826216
                               |
                badeconomy#inc |
           Economy worsened#1  |  -.0957692   .0214299    -4.47   0.000    -.1377827   -.0537557
                               |
                    copsource2 |
       Out-partisan newspaper  |   -.001638   .0102802    -0.16   0.873    -.0217925    .0185165
        Co-partisan newspaper  |   .0284233     .01306     2.18   0.030      .002819    .0540277
                               |
                        female |
                       Female  |  -.0195714    .009245    -2.12   0.034    -.0376963   -.0014464
                               |
                       country |
                        Chile  |  -.0075427   .0090039    -0.84   0.402     -.025195    .0101096
                      Uruguay  |   .0020141   .0091108     0.22   0.825    -.0158477    .0198758
                               |
                         _cons |   .6201845   .0251552    24.65   0.000     .5708675    .6695015
------------------------------------------------------------------------------------------------
(est4 stored)

. eststo: reg vote i.corrupttreat##i.lowealth i.copartisan i.badeconomy##inc i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(16, 4333)       =      64.47
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1140
                                                Root MSE          =      .4269

                                           (Std. Err. adjusted for 4,334 clusters in uniq_id)
---------------------------------------------------------------------------------------------
                            |               Robust
                       vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------------+----------------------------------------------------------------
               corrupttreat |
                    Bribes  |  -.4102308   .0207106   -19.81   0.000    -.4508343   -.3696274
             Bribes common  |  -.4286157   .0207084   -20.70   0.000    -.4692148   -.3880167
           Bribes but jobs  |  -.3515039    .022452   -15.66   0.000    -.3955214   -.3074864
                            |
                   lowealth |
                Low wealth  |  -.1064546    .020212    -5.27   0.000    -.1460804   -.0668287
                            |
      corrupttreat#lowealth |
         Bribes#Low wealth  |   .0965514    .027083     3.57   0.000     .0434548    .1496479
  Bribes common#Low wealth  |   .1112908   .0271219     4.10   0.000      .058118    .1644636
Bribes but jobs#Low wealth  |   .1452486   .0293066     4.96   0.000     .0877928    .2027045
                            |
                 copartisan |
     Co-partisan candidate  |   .0778338   .0189703     4.10   0.000     .0406422    .1150254
                            |
                 badeconomy |
          Economy worsened  |   .0287534   .0129572     2.22   0.027     .0033506    .0541561
                      1.inc |   .0524915   .0153401     3.42   0.001      .022417     .082566
                            |
             badeconomy#inc |
        Economy worsened#1  |  -.0956775   .0214441    -4.46   0.000    -.1377188   -.0536361
                            |
                 copsource2 |
    Out-partisan newspaper  |  -.0014873   .0102901    -0.14   0.885    -.0216611    .0186865
     Co-partisan newspaper  |   .0278001   .0130946     2.12   0.034     .0021281    .0534722
                            |
                     female |
                    Female  |  -.0192977   .0092489    -2.09   0.037    -.0374303   -.0011652
                            |
                    country |
                     Chile  |  -.0069381   .0090106    -0.77   0.441    -.0246036    .0107273
                   Uruguay  |   .0027358   .0091111     0.30   0.764    -.0151265    .0205982
                            |
                      _cons |   .5829475   .0199856    29.17   0.000     .5437656    .6221295
---------------------------------------------------------------------------------------------
(est5 stored)

. esttab, label b(2) se(2) nonumbers nobaselevels nolegend star(* 0.10 ** 0.05 *** 0.01) ///
>         mlabels("Figure 1" "Figure 2" "Figure 2" "Figure 3" "Figure 3")

----------------------------------------------------------------------------------------------------
                         Figure 1        Figure 2        Figure 2        Figure 3        Figure 3   
----------------------------------------------------------------------------------------------------
Bribes                      -0.35***        -0.35***        -0.37***        -0.46***        -0.41***
                           (0.01)          (0.01)          (0.01)          (0.03)          (0.02)   

Bribes common               -0.36***        -0.36***        -0.37***        -0.49***        -0.43***
                           (0.01)          (0.01)          (0.01)          (0.03)          (0.02)   

Bribes but jobs             -0.26***        -0.26***        -0.28***        -0.42***        -0.35***
                           (0.01)          (0.01)          (0.02)          (0.03)          (0.02)   

Co-partisan candid~e         0.08***         0.16***         0.08***         0.08***         0.08***
                           (0.02)          (0.04)          (0.02)          (0.02)          (0.02)   

Economy worsened             0.03**          0.03**          0.03**          0.03**          0.03** 
                           (0.01)          (0.01)          (0.01)          (0.01)          (0.01)   

Experiment: Incumb~1         0.05***         0.05***         0.05***         0.05***         0.05***
                           (0.02)          (0.02)          (0.02)          (0.02)          (0.02)   

Economy worsened #~I        -0.10***        -0.10***        -0.10***        -0.10***        -0.10***
                           (0.02)          (0.02)          (0.02)          (0.02)          (0.02)   

Out-partisan newsp~r        -0.00           -0.00           -0.00           -0.00           -0.00   
                           (0.01)          (0.01)          (0.01)          (0.01)          (0.01)   

Co-partisan newspa~r         0.03**          0.03**          0.02*           0.03**          0.03** 
                           (0.01)          (0.01)          (0.01)          (0.01)          (0.01)   

Female                      -0.02**         -0.02**         -0.02**         -0.02**         -0.02** 
                           (0.01)          (0.01)          (0.01)          (0.01)          (0.01)   

Chile                       -0.01           -0.01           -0.01           -0.01           -0.01   
                           (0.01)          (0.01)          (0.01)          (0.01)          (0.01)   

Uruguay                      0.00            0.00            0.00            0.00            0.00   
                           (0.01)          (0.01)          (0.01)          (0.01)          (0.01)   

Bribes # Co-partis~e                        -0.11**                                                 
                                           (0.06)                                                   

Bribes common # Co~d                        -0.07                                                   
                                           (0.05)                                                   

Bribes but jobs # ~a                        -0.13**                                                 
                                           (0.06)                                                   

High tolerance                                               0.02                                   
                                                           (0.03)                                   

Bribes # High tole~e                                         0.08**                                 
                                                           (0.04)                                   

Bribes common # Hi~e                                         0.00                                   
                                                           (0.04)                                   

Bribes but jobs # ~e                                         0.12***                                
                                                           (0.05)                                   

Low education                                                               -0.12***                
                                                                           (0.02)                   

Bribes # Low educa~n                                                         0.13***                
                                                                           (0.03)                   

Bribes common # Lo~n                                                         0.16***                
                                                                           (0.03)                   

Bribes but jobs # ~n                                                         0.19***                
                                                                           (0.03)                   

Low wealth                                                                                  -0.11***
                                                                                           (0.02)   

Bribes # Low wealth                                                                          0.10***
                                                                                           (0.03)   

Bribes common # Lo~h                                                                         0.11***
                                                                                           (0.03)   

Bribes but jobs # ~h                                                                         0.15***
                                                                                           (0.03)   

Constant                     0.52***         0.52***         0.52***         0.62***         0.58***
                           (0.02)          (0.02)          (0.02)          (0.03)          (0.02)   
----------------------------------------------------------------------------------------------------
Observations                 8668            8668            8566            8668            8668   
----------------------------------------------------------------------------------------------------
Standard errors in parentheses

. 
. 
. *** Table A5: Estimates with OLS and Logit
. eststo clear

. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. eststo ols: reg vote i.corrupttreat i.copartisan i.badeconomy##i.inc i.copsource2 ///
>         i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(12, 4333)       =      82.21
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1103
                                                Root MSE          =      .4277

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3533974    .013401   -26.37   0.000    -.3796702   -.3271247
         Bribes common  |  -.3631275   .0134214   -27.06   0.000    -.3894404   -.3368146
       Bribes but jobs  |   -.264988   .0145094   -18.26   0.000    -.2934338   -.2365422
                        |
             copartisan |
 Co-partisan candidate  |   .0775461   .0188731     4.11   0.000     .0405451     .114547
                        |
             badeconomy |
      Economy worsened  |   .0291838   .0129845     2.25   0.025     .0037275      .05464
                  1.inc |   .0528361   .0153538     3.44   0.001     .0227349    .0829373
                        |
         badeconomy#inc |
    Economy worsened#1  |  -.0968282    .021476    -4.51   0.000    -.1389322   -.0547242
                        |
             copsource2 |
Out-partisan newspaper  |  -.0015192   .0103112    -0.15   0.883    -.0217345    .0186961
 Co-partisan newspaper  |   .0283579   .0130783     2.17   0.030     .0027178    .0539981
                        |
                 female |
                Female  |  -.0191172   .0092638    -2.06   0.039    -.0372791   -.0009553
                        |
                country |
                 Chile  |  -.0083276   .0090087    -0.92   0.355    -.0259894    .0093341
               Uruguay  |   .0021432   .0091276     0.23   0.814    -.0157516    .0200379
                        |
                  _cons |   .5207411   .0163245    31.90   0.000     .4887368    .5527454
-----------------------------------------------------------------------------------------

. logit vote i.corrupttreat i.copartisan i.badeconomy##i.inc i.copsource2 ///
>         i.female i.country [pw=weight1500], cl(uniq_id)

Iteration 0:   log pseudolikelihood = -5020.1719  
Iteration 1:   log pseudolikelihood = -4585.3623  
Iteration 2:   log pseudolikelihood = -4575.7106  
Iteration 3:   log pseudolikelihood = -4575.7062  
Iteration 4:   log pseudolikelihood = -4575.7062  

Logistic regression                             Number of obs     =      8,668
                                                Wald chi2(12)     =     885.03
                                                Prob > chi2       =     0.0000
Log pseudolikelihood = -4575.7062               Pseudo R2         =     0.0885

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -1.655878   .0700114   -23.65   0.000    -1.793098   -1.518659
         Bribes common  |  -1.723508   .0719339   -23.96   0.000    -1.864496    -1.58252
       Bribes but jobs  |  -1.143435   .0662713   -17.25   0.000    -1.273324   -1.013545
                        |
             copartisan |
 Co-partisan candidate  |   .3992877   .0933689     4.28   0.000      .216288    .5822875
                        |
             badeconomy |
      Economy worsened  |   .1603618   .0712024     2.25   0.024     .0208076    .2999159
                  1.inc |   .2834265    .082484     3.44   0.001     .1217608    .4450921
                        |
         badeconomy#inc |
    Economy worsened#1  |  -.5305131   .1178111    -4.50   0.000    -.7614185   -.2996076
                        |
             copsource2 |
Out-partisan newspaper  |  -.0073293   .0571571    -0.13   0.898    -.1193551    .1046965
 Co-partisan newspaper  |   .1534195   .0698748     2.20   0.028     .0164674    .2903717
                        |
                 female |
                Female  |  -.1050914   .0507168    -2.07   0.038    -.2044945   -.0056882
                        |
                country |
                 Chile  |  -.0463768   .0497395    -0.93   0.351    -.1438644    .0511108
               Uruguay  |   .0100864   .0496907     0.20   0.839    -.0873057    .1074785
                        |
                  _cons |   .0672235   .0812664     0.83   0.408    -.0920558    .2265027
-----------------------------------------------------------------------------------------

. eststo mfx: margins, dydx(*) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Pr(vote), predict()
dy/dx w.r.t. : 2.corrupttreat 3.corrupttreat 4.corrupttreat 1.copartisan 1.badeconomy 1.inc 1.copsource2 2.copsource2 1.female 2.country 3.country

-----------------------------------------------------------------------------------------
                        |            Delta-method
                        |      dy/dx   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3532907   .0133869   -26.39   0.000    -.3795285   -.3270529
         Bribes common  |  -.3631235   .0134157   -27.07   0.000    -.3894178   -.3368291
       Bribes but jobs  |  -.2648855   .0145136   -18.25   0.000    -.2933316   -.2364393
                        |
             copartisan |
 Co-partisan candidate  |   .0771802    .018928     4.08   0.000     .0400821    .1142783
                        |
             badeconomy |
      Economy worsened  |  -.0192434   .0073279    -2.63   0.009    -.0336059    -.004881
                  1.inc |   .0047109   .0107321     0.44   0.661    -.0163236    .0257454
                        |
             copsource2 |
Out-partisan newspaper  |  -.0013242   .0103285    -0.13   0.898    -.0215676    .0189193
 Co-partisan newspaper  |   .0285086   .0130597     2.18   0.029     .0029121    .0541052
                        |
                 female |
                Female  |  -.0191899   .0092604    -2.07   0.038    -.0373399   -.0010399
                        |
                country |
                 Chile  |  -.0084364   .0090478    -0.93   0.351    -.0261697    .0092969
               Uruguay  |   .0018532   .0091299     0.20   0.839    -.0160411    .0197476
-----------------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. logit vote i.corrupttreat i.copartisan i.badeconomy##i.inc i.copsource2 ///
>         i.female i.country [pw=weight1500], cl(uniq_id)

Iteration 0:   log pseudolikelihood = -5020.1719  
Iteration 1:   log pseudolikelihood = -4585.3623  
Iteration 2:   log pseudolikelihood = -4575.7106  
Iteration 3:   log pseudolikelihood = -4575.7062  
Iteration 4:   log pseudolikelihood = -4575.7062  

Logistic regression                             Number of obs     =      8,668
                                                Wald chi2(12)     =     885.03
                                                Prob > chi2       =     0.0000
Log pseudolikelihood = -4575.7062               Pseudo R2         =     0.0885

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -1.655878   .0700114   -23.65   0.000    -1.793098   -1.518659
         Bribes common  |  -1.723508   .0719339   -23.96   0.000    -1.864496    -1.58252
       Bribes but jobs  |  -1.143435   .0662713   -17.25   0.000    -1.273324   -1.013545
                        |
             copartisan |
 Co-partisan candidate  |   .3992877   .0933689     4.28   0.000      .216288    .5822875
                        |
             badeconomy |
      Economy worsened  |   .1603618   .0712024     2.25   0.024     .0208076    .2999159
                  1.inc |   .2834265    .082484     3.44   0.001     .1217608    .4450921
                        |
         badeconomy#inc |
    Economy worsened#1  |  -.5305131   .1178111    -4.50   0.000    -.7614185   -.2996076
                        |
             copsource2 |
Out-partisan newspaper  |  -.0073293   .0571571    -0.13   0.898    -.1193551    .1046965
 Co-partisan newspaper  |   .1534195   .0698748     2.20   0.028     .0164674    .2903717
                        |
                 female |
                Female  |  -.1050914   .0507168    -2.07   0.038    -.2044945   -.0056882
                        |
                country |
                 Chile  |  -.0463768   .0497395    -0.93   0.351    -.1438644    .0511108
               Uruguay  |   .0100864   .0496907     0.20   0.839    -.0873057    .1074785
                        |
                  _cons |   .0672235   .0812664     0.83   0.408    -.0920558    .2265027
-----------------------------------------------------------------------------------------

. margins, dydx(badeconomy) at(inc = (0 1))

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Pr(vote), predict()
dy/dx w.r.t. : 1.badeconomy

1._at        : inc             =           0

2._at        : inc             =           1

-------------------------------------------------------------------------------
              |            Delta-method
              |      dy/dx   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------+----------------------------------------------------------------
0.badeconomy  |  (base outcome)
--------------+----------------------------------------------------------------
1.badeconomy  |
          _at |
           1  |   .0292112   .0129631     2.25   0.024      .003804    .0546183
           2  |  -.0678132   .0130218    -5.21   0.000    -.0933355   -.0422909
-------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. margins, dydx(inc) at(badeconomy = 0)

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Pr(vote), predict()
dy/dx w.r.t. : 1.inc
at           : badeconomy      =           0

------------------------------------------------------------------------------
             |            Delta-method
             |      dy/dx   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       1.inc |   .0527323   .0153143     3.44   0.001     .0227169    .0827477
------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. * (manually fix the economy effects in column 2)
. esttab ols mfx, label b(3) se(2) nonumbers nobaselevels nolegend star(* 0.10 ** 0.05 *** 0.01) ///
>         mtitles("OLS" "Logit") drop(_cons)

----------------------------------------------------
                              OLS           Logit   
----------------------------------------------------
Bribes                     -0.353***       -0.353***
                           (0.01)          (0.01)   

Bribes common              -0.363***       -0.363***
                           (0.01)          (0.01)   

Bribes but jobs            -0.265***       -0.265***
                           (0.01)          (0.01)   

Co-partisan candid~e        0.078***        0.077***
                           (0.02)          (0.02)   

Economy worsened            0.029**        -0.019***
                           (0.01)          (0.01)   

Experiment: Incumb~1        0.053***        0.005   
                           (0.02)          (0.01)   

Economy worsened #~I       -0.097***                
                           (0.02)                   

Out-partisan newsp~r       -0.002          -0.001   
                           (0.01)          (0.01)   

Co-partisan newspa~r        0.028**         0.029** 
                           (0.01)          (0.01)   

Female                     -0.019**        -0.019** 
                           (0.01)          (0.01)   

Chile                      -0.008          -0.008   
                           (0.01)          (0.01)   

Uruguay                     0.002           0.002   
                           (0.01)          (0.01)   
----------------------------------------------------
Observations                 8668            8668   
----------------------------------------------------
Standard errors in parentheses

. 
.         
. *** Table A6: Multiple-Comparison Correction for Main Estimates
. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. reg vote i.corrupttreat i.copartisan i.badeconomy##i.inc i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(12, 4333)       =      82.21
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1103
                                                Root MSE          =      .4277

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3533974    .013401   -26.37   0.000    -.3796702   -.3271247
         Bribes common  |  -.3631275   .0134214   -27.06   0.000    -.3894404   -.3368146
       Bribes but jobs  |   -.264988   .0145094   -18.26   0.000    -.2934338   -.2365422
                        |
             copartisan |
 Co-partisan candidate  |   .0775461   .0188731     4.11   0.000     .0405451     .114547
                        |
             badeconomy |
      Economy worsened  |   .0291838   .0129845     2.25   0.025     .0037275      .05464
                  1.inc |   .0528361   .0153538     3.44   0.001     .0227349    .0829373
                        |
         badeconomy#inc |
    Economy worsened#1  |  -.0968282    .021476    -4.51   0.000    -.1389322   -.0547242
                        |
             copsource2 |
Out-partisan newspaper  |  -.0015192   .0103112    -0.15   0.883    -.0217345    .0186961
 Co-partisan newspaper  |   .0283579   .0130783     2.17   0.030     .0027178    .0539981
                        |
                 female |
                Female  |  -.0191172   .0092638    -2.06   0.039    -.0372791   -.0009553
                        |
                country |
                 Chile  |  -.0083276   .0090087    -0.92   0.355    -.0259894    .0093341
               Uruguay  |   .0021432   .0091276     0.23   0.814    -.0157516    .0200379
                        |
                  _cons |   .5207411   .0163245    31.90   0.000     .4887368    .5527454
-----------------------------------------------------------------------------------------

. mat coef = e(b)

. mat pval = r(table)

. local C = colsof(pval)

. mat pval = pval[4..4,1..`C']

. mat coef = coef', pval'

. scalar R = rowsof(coef)-1

. mat coef = coef[1..R,1..2]

. clear

. svmat double coef
number of observations will be reset to 22
Press any key to continue, or Break to abort
number of observations (_N) was 0, now 22

. gen names = ""
(22 missing values generated)

. local names : rownames coef

. local k = 1

. foreach i of local names {
  2. replace names = "`i'" in `k'
  3. local k = `k' + 1
  4. }
variable names was str1 now str15
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
variable names was str15 now str20
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)
(1 real change made)

. egen base = rowtotal(coef*)

. drop if base == 0
(10 observations deleted)

. drop base

. rename coef2 pval

. gen orig_order = _n

. sort pval

. gen pval_order = _n if pval ~= .

. sum pval_order

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
  pval_order |         12         6.5    3.605551          1         12

. local max = r(max)

. gen bh1 = (pval_order/`max')*0.05

. gen bh2 = (pval < bh1)

. gen orig = (pval < .05)

. gen fail = (orig ~= bh2)

. sort orig_order

. drop in 11/12
(2 observations deleted)

. mkmat coef1 pval orig bh1 bh2, mat(ests)

. frmttable, statmat(ests) substat(0) sdec(3,3,0,3,0) ///
>         rtitles("Bribes" \ "Bribes common" \ "Bribes but jobs" \ ///
>         "Co-partisan candidate" \ "Worsened economy" \ "Incumbent profile" \ "Worsened economy $\times$ incumbent" \ ///
>         "Out-partisan source" \ "Co-partisan source" \ "Female") ///
>         ctitles("","Point","Original","Sig. at","B-H \$ p$-value","Sig. at" \ ///
>         "","Estimate","$ p$-value","$ p<.05$","threshold","B-H $ p<.05$")

                                            -----------------------------------------------------------------------------------------------------
                                                                                    Point     Original   Sig. at   B-H $ p$-value    Sig. at    
                                                                                   Estimate  $ p$-value  $ p<.05$    threshold     B-H $ p<.05$ 
                                            -----------------------------------------------------------------------------------------------------
                                             Bribes                                 -0.353     0.000        1          0.008            1       
                                             Bribes common                          -0.363     0.000        1          0.004            1       
                                             Bribes but jobs                        -0.265     0.000        1          0.013            1       
                                             Co-partisan candidate                  0.078      0.000        1          0.021            1       
                                             Worsened economy                       0.029      0.025        1          0.029            1       
                                             Incumbent profile                      0.053      0.001        1          0.025            1       
                                             Worsened economy $\times$ incumbent    -0.097     0.000        1          0.017            1       
                                             Out-partisan source                    -0.002     0.883        0          0.050            0       
                                             Co-partisan source                     0.028      0.030        1          0.033            1       
                                             Female                                 -0.019     0.039        1          0.038            0       
                                            -----------------------------------------------------------------------------------------------------


. frmttable, clear

. 
.         
. *** Figure A2: Treatment Effects, by Country
. cap estimates store clear

. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. reg vote i.corrupttreat badeconomy##arias i.copartisan i.copsource2 i.female if country == 1 [pw=weight1500], cl(uniq_id)
(sum of wgt is 2,719.24082040787)

Linear regression                               Number of obs     =      2,770
                                                F(10, 1384)       =      31.75
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1105
                                                Root MSE          =      .4283

                                       (Std. Err. adjusted for 1,385 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3420875   .0240708   -14.21   0.000    -.3893067   -.2948683
         Bribes common  |  -.3789938   .0233447   -16.23   0.000    -.4247885    -.333199
       Bribes but jobs  |  -.2419911   .0260179    -9.30   0.000    -.2930298   -.1909523
                        |
             badeconomy |
      Economy worsened  |  -.0591574   .0225562    -2.62   0.009    -.1034055   -.0149094
                1.arias |  -.0016708   .0276777    -0.06   0.952    -.0559656    .0526239
                        |
       badeconomy#arias |
    Economy worsened#1  |   .0608175   .0383688     1.59   0.113    -.0144497    .1360847
                        |
             copartisan |
 Co-partisan candidate  |  -.0605388   .0430163    -1.41   0.160    -.1449229    .0238454
                        |
             copsource2 |
Out-partisan newspaper  |   .0054618   .0183693     0.30   0.766    -.0305729    .0414965
 Co-partisan newspaper  |   .0290311   .0229317     1.27   0.206    -.0159536    .0740158
                        |
                 female |
                Female  |  -.0148531    .016534    -0.90   0.369    -.0472875    .0175814
                  _cons |   .5471886    .026967    20.29   0.000     .4942881    .6000892
-----------------------------------------------------------------------------------------

. estimates store arg

. reg vote i.corrupttreat badeconomy##arias i.copartisan i.copsource2 i.female if country == 2 [pw=weight1500], cl(uniq_id)
(sum of wgt is 2,832.0000140667)

Linear regression                               Number of obs     =      3,068
                                                F(10, 1533)       =      23.67
                                                Prob > F          =     0.0000
                                                R-squared         =     0.0785
                                                Root MSE          =     .43086

                                       (Std. Err. adjusted for 1,534 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.2972877   .0224701   -13.23   0.000     -.341363   -.2532124
         Bribes common  |  -.3133588   .0229247   -13.67   0.000    -.3583259   -.2683916
       Bribes but jobs  |  -.2028376   .0244836    -8.28   0.000    -.2508625   -.1548127
                        |
             badeconomy |
      Economy worsened  |  -.0580649   .0222628    -2.61   0.009    -.1017337   -.0143961
                1.arias |  -.0435335   .0259614    -1.68   0.094    -.0944571      .00739
                        |
       badeconomy#arias |
    Economy worsened#1  |   .0806765   .0361357     2.23   0.026     .0097959     .151557
                        |
             copartisan |
 Co-partisan candidate  |   .0119142   .0378624     0.31   0.753    -.0623534    .0861818
                        |
             copsource2 |
Out-partisan newspaper  |   -.003104   .0174669    -0.18   0.859    -.0373655    .0311575
 Co-partisan newspaper  |   .0174408   .0225567     0.77   0.440    -.0268046    .0616861
                        |
                 female |
                Female  |    .003689   .0153465     0.24   0.810    -.0264134    .0337914
                  _cons |   .5110256   .0260848    19.59   0.000       .45986    .5621912
-----------------------------------------------------------------------------------------

. estimates store chi

. reg vote i.corrupttreat badeconomy##arias i.copartisan i.copsource2 i.female if country == 3 [pw=weight1500], cl(uniq_id)
(sum of wgt is 2,803.83097589016)

Linear regression                               Number of obs     =      2,830
                                                F(10, 1414)       =      52.60
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1599
                                                Root MSE          =     .42041

                                       (Std. Err. adjusted for 1,415 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.4172786   .0230435   -18.11   0.000    -.4624816   -.3720755
         Bribes common  |  -.3964025    .023328   -16.99   0.000    -.4421637   -.3506413
       Bribes but jobs  |  -.3482136   .0245432   -14.19   0.000    -.3963586   -.3000686
                        |
             badeconomy |
      Economy worsened  |  -.0816348    .022799    -3.58   0.000    -.1263583   -.0369113
                1.arias |  -.1089463   .0259413    -4.20   0.000    -.1598339   -.0580588
                        |
       badeconomy#arias |
    Economy worsened#1  |   .1424811   .0368729     3.86   0.000     .0701495    .2148126
                        |
             copartisan |
 Co-partisan candidate  |   .1320338   .0241661     5.46   0.000     .0846285    .1794391
                        |
             copsource2 |
Out-partisan newspaper  |  -.0085714   .0176807    -0.48   0.628    -.0432548    .0261119
 Co-partisan newspaper  |   .0376229   .0222275     1.69   0.091    -.0059794    .0812253
                        |
                 female |
                Female  |  -.0475972   .0161346    -2.95   0.003    -.0792476   -.0159468
                  _cons |   .6528092   .0262368    24.88   0.000      .601342    .7042764
-----------------------------------------------------------------------------------------

. estimates store uru

. coefplot (arg, label(Argentina) msymbol(O) msize(vsmall) offset(0.27)) ///
>         (chi, label(Chile) msymbol(T) msize(vsmall) offset(0)) ///
>         (uru, label(Uruguay) msymbol(S) msize(vsmall) offset(-0.27)), ///
>         scheme(plotplain) xline(0, lcolor(black) lpattern(solid) lwidth(vthin)) baselevels ///
>         graphregion(color(white)) drop(_cons 0.arias 0.badeconomy#0.arias 0.badeconomy#1.arias 1.badeconomy#0.arias ///
>                 1.arias 1.badeconomy#1.arias) legend(ring(0) position(1)) ///
>                 xtitle("Effect on Pr(Voting for candidate)") coeflabels(, labsize(small)) grid(glcolor(gs2)) ///
>                 headings(1.corrupttreat = "{bf:Corruption}" ///
>                         0.badeconomy = "{bf:Economy} (incumbent only)" ///
>                         0.female = "{bf: Gender}" ///
>                         0.copartisan = "{bf:Partisanship}" ///
>                         0.copsource2 = "{bf:Information source}")

. 
. 
. *** Table A7: Corruption Treatment Effects, by Country
. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. forval i = 2/4 {
  2.         mat b`i' = J(1,6,.)
  3.         mat bp`i' = J(1,6,.)
  4.         mat d`i' = J(1,6,.)
  5.         mat dp`i' = J(1,6,.)
  6.         }

. forval a = 2/4 {
  2.         * effects
.         qui reg vote i.corrupttreat##i.country badeconomy##arias i.copartisan i.copsource2 i.female [pw=weight1500], cl(uniq_id)
  3.         margins, dydx(`a'.corrupttreat) at(country = (1 2 3)) post
  4.         mat tab = r(table)
  5.         local j 1 3 5
  6.         local k 2 4 6
  7.         forval i = 4/6 {
  8.                 local m = `i'-3
  9.                 local n : word `m' of `j'
 10.                 local p : word `m' of `k'
 11.                 mat b`a'[1,`n'] = tab[1,`i']
 12.                 mat b`a'[1,`p'] = tab[2,`i']
 13.                 mat bp`a'[1,`n'] = tab[4,`i']
 14.                 }
 15.         * differences across countries
.         qui reg vote i.corrupttreat##i.country badeconomy##arias i.copartisan i.copsource2 i.female [pw=weight1500], cl(uniq_id)
 16.         margins, dydx(`a'.corrupttreat) at(country = (1 2 3)) post
 17.         * Argentina vs. Chile
.         lincom _b[`a'.corrupttreat:1._at]-_b[`a'.corrupttreat:2._at]
 18.         mat d`a'[1,1] = r(estimate)
 19.         mat d`a'[1,2] = r(se)
 20.         mat dp`a'[1,1] = r(p)
 21.         * Argentina vs. Uruguay
.         lincom _b[`a'.corrupttreat:1._at]-_b[`a'.corrupttreat:3._at]
 22.         mat d`a'[1,3] = r(estimate)
 23.         mat d`a'[1,4] = r(se)
 24.         mat dp`a'[1,3] = r(p)
 25.         * Chile vs. Uruguay
.         lincom _b[`a'.corrupttreat:2._at]-_b[`a'.corrupttreat:3._at]
 26.         mat d`a'[1,5] = r(estimate)
 27.         mat d`a'[1,6] = r(se)
 28.         mat dp`a'[1,5] = r(p)
 29.         }

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat

1._at        : country         =           1

2._at        : country         =           2

3._at        : country         =           3

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |  -.3436399   .0240434   -14.29   0.000    -.3907773   -.2965025
             2  |  -.2972717     .02249   -13.22   0.000    -.3413636   -.2531797
             3  |  -.4191868   .0230237   -18.21   0.000    -.4643251   -.3740485
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat

1._at        : country         =           1

2._at        : country         =           2

3._at        : country         =           3

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |  -.3436399   .0240434   -14.29   0.000    -.3907773   -.2965025
             2  |  -.2972717     .02249   -13.22   0.000    -.3413636   -.2531797
             3  |  -.4191868   .0230237   -18.21   0.000    -.4643251   -.3740485
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

 ( 1)  [2.corrupttreat]1bn._at - [2.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |  -.0463682   .0329232    -1.41   0.159    -.1109145    .0181781
------------------------------------------------------------------------------

 ( 1)  [2.corrupttreat]1bn._at - [2.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |    .075547   .0333003     2.27   0.023     .0102613    .1408327
------------------------------------------------------------------------------

 ( 1)  [2.corrupttreat]2._at - [2.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .1219152   .0321917     3.79   0.000     .0588029    .1850274
------------------------------------------------------------------------------

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 3.corrupttreat

1._at        : country         =           1

2._at        : country         =           2

3._at        : country         =           3

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
3.corrupttreat  |
            _at |
             1  |  -.3798506   .0233583   -16.26   0.000    -.4256448   -.3340563
             2  |  -.3129975   .0229037   -13.67   0.000    -.3579004   -.2680945
             3  |  -.3967184   .0232997   -17.03   0.000    -.4423978    -.351039
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 3.corrupttreat

1._at        : country         =           1

2._at        : country         =           2

3._at        : country         =           3

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
3.corrupttreat  |
            _at |
             1  |  -.3798506   .0233583   -16.26   0.000    -.4256448   -.3340563
             2  |  -.3129975   .0229037   -13.67   0.000    -.3579004   -.2680945
             3  |  -.3967184   .0232997   -17.03   0.000    -.4423978    -.351039
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

 ( 1)  [3.corrupttreat]1bn._at - [3.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |  -.0668531   .0327059    -2.04   0.041    -.1309734   -.0027328
------------------------------------------------------------------------------

 ( 1)  [3.corrupttreat]1bn._at - [3.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0168678    .032991     0.51   0.609    -.0478115    .0815471
------------------------------------------------------------------------------

 ( 1)  [3.corrupttreat]2._at - [3.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0837209   .0326628     2.56   0.010     .0196851    .1477568
------------------------------------------------------------------------------

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 4.corrupttreat

1._at        : country         =           1

2._at        : country         =           2

3._at        : country         =           3

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
4.corrupttreat  |
            _at |
             1  |   -.242106   .0261007    -9.28   0.000    -.2932768   -.1909352
             2  |  -.2021718   .0245084    -8.25   0.000    -.2502208   -.1541227
             3  |  -.3498301   .0244954   -14.28   0.000    -.3978537   -.3018065
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 4.corrupttreat

1._at        : country         =           1

2._at        : country         =           2

3._at        : country         =           3

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
4.corrupttreat  |
            _at |
             1  |   -.242106   .0261007    -9.28   0.000    -.2932768   -.1909352
             2  |  -.2021718   .0245084    -8.25   0.000    -.2502208   -.1541227
             3  |  -.3498301   .0244954   -14.28   0.000    -.3978537   -.3018065
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

 ( 1)  [4.corrupttreat]1bn._at - [4.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |  -.0399342   .0357983    -1.12   0.265    -.1101173    .0302488
------------------------------------------------------------------------------

 ( 1)  [4.corrupttreat]1bn._at - [4.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .1077241   .0357761     3.01   0.003     .0375847    .1778635
------------------------------------------------------------------------------

 ( 1)  [4.corrupttreat]2._at - [4.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .1476583   .0346309     4.26   0.000     .0797641    .2155526
------------------------------------------------------------------------------

. * differences across treatments within country
. qui reg vote i.corrupttreat##i.country badeconomy##arias i.copartisan i.copsource2 i.female [pw=weight1500], cl(uniq_id)

. margins, dydx(corrupttreat) at(country = (1 2 3)) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat 3.corrupttreat 4.corrupttreat

1._at        : country         =           1

2._at        : country         =           2

3._at        : country         =           3

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |  -.3436399   .0240434   -14.29   0.000    -.3907773   -.2965025
             2  |  -.2972717     .02249   -13.22   0.000    -.3413636   -.2531797
             3  |  -.4191868   .0230237   -18.21   0.000    -.4643251   -.3740485
----------------+----------------------------------------------------------------
3.corrupttreat  |
            _at |
             1  |  -.3798506   .0233583   -16.26   0.000    -.4256448   -.3340563
             2  |  -.3129975   .0229037   -13.67   0.000    -.3579004   -.2680945
             3  |  -.3967184   .0232997   -17.03   0.000    -.4423978    -.351039
----------------+----------------------------------------------------------------
4.corrupttreat  |
            _at |
             1  |   -.242106   .0261007    -9.28   0.000    -.2932768   -.1909352
             2  |  -.2021718   .0245084    -8.25   0.000    -.2502208   -.1541227
             3  |  -.3498301   .0244954   -14.28   0.000    -.3978537   -.3018065
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. forval i = 1/2 {
  2.         mat db`i' = J(1,6,.)
  3.         mat dbp`i' = J(1,6,.)
  4.         mat dd`i' = J(1,6,.)
  5.         mat ddp`i' = J(1,6,.)
  6.         }

. local j 1 3 5

. local k 2 4 6

. forval i = 1/3 {
  2.         local n : word `i' of `j'
  3.         local p : word `i' of `k'
  4.         * bribes common vs. bribes
.         lincom _b[3.corrupttreat:`i'._at]-_b[2.corrupttreat:`i'._at]
  5.         mat db1[1,`n'] = r(estimate)
  6.         mat db1[1,`p'] = r(se)
  7.         mat dbp1[1,`n'] = r(p)
  8.         * bribes but jobs vs. bribes
.         lincom _b[4.corrupttreat:`i'._at]-_b[2.corrupttreat:`i'._at]
  9.         mat db2[1,`n'] = r(estimate)
 10.         mat db2[1,`p'] = r(se)
 11.         mat dbp2[1,`n'] = r(p)
 12.         }

 ( 1)  - [2.corrupttreat]1bn._at + [3.corrupttreat]1bn._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |  -.0362107   .0200269    -1.81   0.071    -.0754738    .0030523
------------------------------------------------------------------------------

 ( 1)  - [2.corrupttreat]1bn._at + [4.corrupttreat]1bn._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .1015339   .0228964     4.43   0.000     .0566453    .1464224
------------------------------------------------------------------------------

 ( 1)  - [2.corrupttreat]2._at + [3.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |  -.0157258   .0195449    -0.80   0.421    -.0540439    .0225923
------------------------------------------------------------------------------

 ( 1)  - [2.corrupttreat]2._at + [4.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0950999   .0217382     4.37   0.000     .0524819    .1377179
------------------------------------------------------------------------------

 ( 1)  - [2.corrupttreat]3._at + [3.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0224684   .0201612     1.11   0.265    -.0170579    .0619948
------------------------------------------------------------------------------

 ( 1)  - [2.corrupttreat]3._at + [4.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0693567   .0216133     3.21   0.001     .0269836    .1117298
------------------------------------------------------------------------------

. * differences in differences: across treatments and countries
. * bribes common vs. bribes
. local j 1 3 5

. local k 2 4 6

. local x 1 1 2

. local y 2 3 3

. forval i = 1/3 {
  2.         local n : word `i' of `j'
  3.         local p : word `i' of `k'
  4.         local x1 : word `i' of `x'
  5.         local y1 : word `i' of `y'
  6.         * bribes common vs. bribes
.         lincom _b[3.corrupttreat:`x1'._at]-_b[2.corrupttreat:`x1'._at]-(_b[3.corrupttreat:`y1'._at]-_b[2.corrupttreat:`y1'._at])
  7.         mat dd1[1,`n'] = r(estimate)
  8.         mat dd1[1,`p'] = r(se)
  9.         mat ddp1[1,`n'] = r(p)
 10.         * bribes but jobs vs. bribes
.         lincom _b[4.corrupttreat:`x1'._at]-_b[2.corrupttreat:`x1'._at]-(_b[4.corrupttreat:`y1'._at]-_b[2.corrupttreat:`y1'._at])
 11.         mat dd2[1,`n'] = r(estimate)
 12.         mat dd2[1,`p'] = r(se)
 13.         mat ddp2[1,`n'] = r(p)
 14.         }

 ( 1)  - [2.corrupttreat]1bn._at + [2.corrupttreat]2._at + [3.corrupttreat]1bn._at - [3.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |  -.0204849   .0279854    -0.73   0.464    -.0753506    .0343808
------------------------------------------------------------------------------

 ( 1)  - [2.corrupttreat]1bn._at + [2.corrupttreat]2._at + [4.corrupttreat]1bn._at - [4.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |    .006434   .0315737     0.20   0.839    -.0554665    .0683345
------------------------------------------------------------------------------

 ( 1)  - [2.corrupttreat]1bn._at + [2.corrupttreat]3._at + [3.corrupttreat]1bn._at - [3.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |  -.0586791   .0284101    -2.07   0.039    -.1143774   -.0029809
------------------------------------------------------------------------------

 ( 1)  - [2.corrupttreat]1bn._at + [2.corrupttreat]3._at + [4.corrupttreat]1bn._at - [4.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0321772   .0314827     1.02   0.307    -.0295451    .0938994
------------------------------------------------------------------------------

 ( 1)  - [2.corrupttreat]2._at + [2.corrupttreat]3._at + [3.corrupttreat]2._at - [3.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |  -.0381942   .0280781    -1.36   0.174    -.0932417    .0168532
------------------------------------------------------------------------------

 ( 1)  - [2.corrupttreat]2._at + [2.corrupttreat]3._at + [4.corrupttreat]2._at - [4.corrupttreat]3._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0257432   .0306566     0.84   0.401    -.0343594    .0858457
------------------------------------------------------------------------------

. * putting all results together into a table
. mat t = J(1,12,.)

. mat top1 = b2 \ b3 \ b4 

. mat top2 = d2 \ d3 \ d4

. mat top = top1, top2

. mat top = t \ top

. mat b = J(1,12,.)

. mat bot1 = db1 \ db2

. mat bot2 = dd1 \ dd2 

. mat bot = bot1, bot2

. mat bot = b \ bot

. mat all = top \ bot

. mat ptop1 = bp2 \ bp3 \ bp4

. mat ptop2 = dp2 \ dp3 \ dp4

. mat ptop = ptop1, ptop2

. mat ptop = t \ ptop

. mat pbot1 = dbp1 \ dbp2

. mat pbot2 = ddp1 \ ddp2

. mat pbot = pbot1, pbot2

. mat pbot = b \ pbot

. mat pval = ptop \ pbot

. mat stars = J(8,12,0)

. local cols = colsof(pval)

. local rows = rowsof(pval)

. forval i = 1/`rows' {
  2.         forval j = 1/`cols' {
  3.                 local pval = pval[`i',`j']
  4.                 if `pval' <= .10 & `pval' > .05 {
  5.                         matrix stars[`i',`j'] = 1
  6.                         }
  7.                 if `pval' <= .05 & `pval' > .01 {
  8.                         matrix stars[`i',`j'] = 2 
  9.                         }
 10.                 if `pval' <= .01 {
 11.                         matrix stars[`i',`j'] = 3
 12.                 }
 13.         }
 14. }

. frmttable, statmat(all) substat(1) sdec(2) annotate(stars) asymbol(*,**,***) ///
>         statfont(fs11) hlines(1001{0}1) vlines(00001{0}) ///
>         rtitles("\textbf{Corruption treatments}" \ "" \ "\hspace{1em} Bribes" \ "" \ ///
>                         "\hspace{1em} Bribes common" \ "" \ "\hspace{1em} Bribes but jobs" \ "" \ ///
>                         "\textbf{Diff. in corruption treatments}" \ "" \ ///
>                         "\hspace{1em} Bribes common vs. bribes" \ "" \ ///
>                         "\hspace{1em} Bribes but jobs vs. bribes" \ "") ///
>         ctitles("","By country","","","Diff. across countries","","" \ ///
>                         "","Argentina","Chile","Uruguay","Argentina","Argentina","Chile" \ ///
>                         "","","","","vs. Chile","vs. Uruguay","vs. Uruguay") ///
>         multicol(1,2,3;1,5,3) noblankrows

                                -----------------------------------------------------------------------------------------------------------------------------
                                                                           By country                      Diff. across countries                           
                                                                           Argentina    Chile    Uruguay         Argentina          Argentina      Chile    
                                                                                                                 vs. Chile         vs. Uruguay  vs. Uruguay 
                                -----------------------------------------------------------------------------------------------------------------------------
                                 \textbf{Corruption treatments}                                                                                             
                                 \hspace{1em} Bribes                        -0.34***   -0.30***  -0.42***          -0.05             0.08**       0.12***   
                                                                             (0.02)     (0.02)    (0.02)           (0.03)            (0.03)       (0.03)    
                                 \hspace{1em} Bribes common                 -0.38***   -0.31***  -0.40***         -0.07**             0.02        0.08**    
                                                                             (0.02)     (0.02)    (0.02)           (0.03)            (0.03)       (0.03)    
                                 \hspace{1em} Bribes but jobs               -0.24***   -0.20***  -0.35***          -0.04             0.11***      0.15***   
                                                                             (0.03)     (0.02)    (0.02)           (0.04)            (0.04)       (0.03)    
                                 \textbf{Diff. in corruption treatments}                                                                                    
                                 \hspace{1em} Bribes common vs. bribes       -0.04*     -0.02      0.02            -0.02             -0.06**       -0.04    
                                                                             (0.02)     (0.02)    (0.02)           (0.03)            (0.03)       (0.03)    
                                 \hspace{1em} Bribes but jobs vs. bribes    0.10***    0.10***   0.07***            0.01              0.03         0.03     
                                                                             (0.02)     (0.02)    (0.02)           (0.03)            (0.03)       (0.03)    
                                -----------------------------------------------------------------------------------------------------------------------------


.         
.         
. *** Figures A3 and A4
. * define program that collects results
. cap prog drop bycorr

. prog def bycorr
  1.         args m /* m = moderator */
  2.         use analysis-data, clear
  3.         mat ests = J(10,3,.)
  4.         mat diffs = J(3,1,.)
  5.         reg vote i.corrupttreat##i.`m' badeconomy##arias i.copartisan i.copsource2 i.female i.country ///
>                 [pw=weight1500], cl(uniq_id)
  6.         margins, dydx(corrupttreat) at(`m' = (0 1)) post
  7.         mat tab = r(table)
  8.         local j 2 3 5 6 8 9
  9.         local k 3 4 5 6 7 8
 10.         forval i = 1/6 {
 11.                 loc m : word `i' of `j'
 12.                 loc n : word `i' of `k'
 13.                 mat ests[`m',1] = tab[1,`n']
 14.                 mat ests[`m',2] = tab[5,`n']
 15.                 mat ests[`m',3] = tab[6,`n']
 16.                 }
 17.         forval i = 1/3 {
 18.                 loc a = `i'+1
 19.                 lincom _b[`a'.corrupttreat:2._at]-_b[`a'.corrupttreat:1._at]
 20.                 mat diffs[`i',1] = r(estimate)
 21.                 }
 22. end

. 
. *** Figure A3: Corruption Treatment Effects, by Respondents’ Corruption Perception
. bycorr corperchi
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)
(sum of wgt is 8,114.34003412724)

Linear regression                               Number of obs     =      8,420
                                                F(16, 4209)       =      64.09
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1136
                                                Root MSE          =     .42851

                                       (Std. Err. adjusted for 4,210 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3643792    .015483   -23.53   0.000     -.394734   -.3340243
         Bribes common  |  -.3787202   .0154453   -24.52   0.000    -.4090012   -.3484391
       Bribes but jobs  |  -.2803104   .0165695   -16.92   0.000    -.3127954   -.2478255
                        |
            1.corperchi |  -.0826963   .0247851    -3.34   0.001    -.1312881   -.0341046
                        |
 corrupttreat#corperchi |
              Bribes#1  |   .0305002   .0319239     0.96   0.339    -.0320875    .0930878
       Bribes common#1  |   .0518413   .0323998     1.60   0.110    -.0116794     .115362
     Bribes but jobs#1  |   .0643406    .035787     1.80   0.072    -.0058208    .1345021
                        |
             badeconomy |
      Economy worsened  |  -.0660495   .0132469    -4.99   0.000    -.0920205   -.0400785
                1.arias |  -.0538813   .0156572    -3.44   0.001    -.0845777   -.0231849
                        |
       badeconomy#arias |
    Economy worsened#1  |   .0973391   .0219081     4.44   0.000     .0543876    .1402905
                        |
             copartisan |
 Co-partisan candidate  |   .0734614    .019238     3.82   0.000     .0357448     .111178
                        |
             copsource2 |
Out-partisan newspaper  |   .0007327   .0104883     0.07   0.944      -.01983    .0212954
 Co-partisan newspaper  |   .0283879   .0132299     2.15   0.032     .0024503    .0543255
                        |
                 female |
                Female  |  -.0152412   .0094234    -1.62   0.106     -.033716    .0032336
                        |
                country |
                 Chile  |  -.0066293   .0090568    -0.73   0.464    -.0243854    .0111268
               Uruguay  |  -.0018105   .0093236    -0.19   0.846    -.0200897    .0164688
                        |
                  _cons |   .5950594    .017116    34.77   0.000     .5615029    .6286158
-----------------------------------------------------------------------------------------

Average marginal effects                        Number of obs     =      8,420
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat 3.corrupttreat 4.corrupttreat

1._at        : corperchi       =           0

2._at        : corperchi       =           1

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |  -.3643792    .015483   -23.53   0.000     -.394734   -.3340243
             2  |   -.333879    .027921   -11.96   0.000    -.3886188   -.2791392
----------------+----------------------------------------------------------------
3.corrupttreat  |
            _at |
             1  |  -.3787202   .0154453   -24.52   0.000    -.4090012   -.3484391
             2  |  -.3268788   .0285005   -11.47   0.000    -.3827549   -.2710028
----------------+----------------------------------------------------------------
4.corrupttreat  |
            _at |
             1  |  -.2803104   .0165695   -16.92   0.000    -.3127954   -.2478255
             2  |  -.2159698   .0317417    -6.80   0.000    -.2782004   -.1537393
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

 ( 1)  - [2.corrupttreat]1bn._at + [2.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0305002   .0319239     0.96   0.339    -.0320875    .0930878
------------------------------------------------------------------------------

 ( 1)  - [3.corrupttreat]1bn._at + [3.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0518413   .0323998     1.60   0.110    -.0116794     .115362
------------------------------------------------------------------------------

 ( 1)  - [4.corrupttreat]1bn._at + [4.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0643406    .035787     1.80   0.072    -.0058208    .1345021
------------------------------------------------------------------------------

. clear

. svmat ests
number of observations will be reset to 10
Press any key to continue, or Break to abort
number of observations (_N) was 0, now 10

. gen n = .
(10 missing values generated)

. for any 1.1 1.5 2 3.1 3.5 4 5.1 5.5 6 6.2 \ any 1 2 3 4 5 6 7 8 9 10: replace n = X in Y

->  replace n = 1.1 in 1
(1 real change made)

->  replace n = 1.5 in 2
(1 real change made)

->  replace n = 2 in 3
(1 real change made)

->  replace n = 3.1 in 4
(1 real change made)

->  replace n = 3.5 in 5
(1 real change made)

->  replace n = 4 in 6
(1 real change made)

->  replace n = 5.1 in 7
(1 real change made)

->  replace n = 5.5 in 8
(1 real change made)

->  replace n = 6 in 9
(1 real change made)

->  replace n = 6.2 in 10
(1 real change made)

. svmat diffs

. gen n2 = .
(10 missing values generated)

. for any 1.75 3.75 5.75 \ any 1 2 3: replace n2 = X in Y

->  replace n2 = 1.75 in 1
(1 real change made)

->  replace n2 = 3.75 in 2
(1 real change made)

->  replace n2 = 5.75 in 3
(1 real change made)

. tostring diffs1, gen(diffs_) force format(%9.3f)
diffs_ generated as str5
diffs_ was forced to string; some loss of information

. gen stars = "" in 1
(10 missing values generated)

. replace stars = "" in 2
(0 real changes made)

. replace stars = "*" in 3
(1 real change made)

. egen diffs = concat(diffs_ stars)

. gen t = "Diff.: " in 1/3
(7 missing values generated)

. egen lab = concat(t diffs)

. forval i = 1/3 {
  2.         loc l`i' = lab in `i'
  3.         di "`l`i''"
  4.         }
Diff.: 0.031
Diff.: 0.052
Diff.: 0.064*

. twoway (rcap ests2 ests3 n, msize(0) lcol(black) horizontal) ///
>                 (scatter n ests1, mcol(black) msize(medium) msymbol(O)), ///
>                 legend(off) scheme(plotplain) ///
>                 ylabel(1.1 "{bf:Bribes}" 1.5 "Low corr. perception" 2 "High corr. perception" ///
>                                 3.1 "{bf:Bribes common}" 3.5 "Low corr. perception" ///
>                                 4 "High corr. perception" 5.1 "{bf:Bribes but jobs}" ///
>                                 5.5 "Low corr. perception" 6 "High corr. perception", ///
>                                 labgap(6pt) noticks nogrid angle(0) labsize(small)) ytitle("") ///
>                 xscale(range(-0.45(.1).05)) xlabel(#6, glcolor(gs2)) xtitle("Effect on Pr(Voting for candidate)") ///
>                 yscale(reverse )  ytick(1.5 2 3.5 4 5.5 6, grid glcolor(gs2)) ///
>                 xline(0, lcolor(black) lpattern(dash) lwidth(vthin))  ///
>                 text(1.71 -.24 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(small) color(black)) ///
>                 text(1.75 -.19 "`l1'", size(small)) ///
>                 text(3.71 -.23 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(small) color(black)) ///
>                 text(3.75 -.18 "`l2'", size(small)) ///
>                 text(5.73 -.13 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(small) color(black)) ///
>                 text(5.75 -.08 "`l3'", size(small)) 

. 
. 
. *** Figure A4: Corruption Treatment Effects, by Respondents’ Bribe Experience
. bycorr anybribe
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(16, 4333)       =      62.44
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1111
                                                Root MSE          =      .4276

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3616029   .0141248   -25.60   0.000    -.3892948   -.3339111
         Bribes common  |  -.3651851   .0142926   -25.55   0.000    -.3932059   -.3371643
       Bribes but jobs  |  -.2731484   .0152684   -17.89   0.000    -.3030824   -.2432144
                        |
             1.anybribe |  -.0229176   .0311974    -0.73   0.463    -.0840804    .0382453
                        |
  corrupttreat#anybribe |
              Bribes#1  |   .0841315   .0449776     1.87   0.061    -.0040477    .1723107
       Bribes common#1  |   .0178566   .0414365     0.43   0.667    -.0633802    .0990934
     Bribes but jobs#1  |   .0813212   .0494292     1.65   0.100    -.0155853    .1782277
                        |
             badeconomy |
      Economy worsened  |  -.0678558   .0130148    -5.21   0.000    -.0933715   -.0423401
                1.arias |  -.0532114   .0153559    -3.47   0.001    -.0833168    -.023106
                        |
       badeconomy#arias |
    Economy worsened#1  |    .097094   .0214671     4.52   0.000     .0550076    .1391804
                        |
             copartisan |
 Co-partisan candidate  |   .0781239   .0188912     4.14   0.000     .0410876    .1151603
                        |
             copsource2 |
Out-partisan newspaper  |  -.0017445   .0103099    -0.17   0.866    -.0219572    .0184682
 Co-partisan newspaper  |   .0284368   .0130824     2.17   0.030     .0027886     .054085
                        |
                 female |
                Female  |  -.0190324   .0092642    -2.05   0.040    -.0371949   -.0008699
                        |
                country |
                 Chile  |  -.0061784   .0090776    -0.68   0.496    -.0239752    .0116183
               Uruguay  |    .004111   .0092079     0.45   0.655    -.0139411    .0221631
                        |
                  _cons |   .5749991   .0165241    34.80   0.000     .5426035    .6073947
-----------------------------------------------------------------------------------------

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat 3.corrupttreat 4.corrupttreat

1._at        : anybribe        =           0

2._at        : anybribe        =           1

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |  -.3616029   .0141248   -25.60   0.000    -.3892948   -.3339111
             2  |  -.2774714   .0427175    -6.50   0.000    -.3612196   -.1937232
----------------+----------------------------------------------------------------
3.corrupttreat  |
            _at |
             1  |  -.3651851   .0142926   -25.55   0.000    -.3932059   -.3371643
             2  |  -.3473285   .0389056    -8.93   0.000    -.4236034   -.2710536
----------------+----------------------------------------------------------------
4.corrupttreat  |
            _at |
             1  |  -.2731484   .0152684   -17.89   0.000    -.3030824   -.2432144
             2  |  -.1918272   .0470017    -4.08   0.000    -.2839747   -.0996798
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

 ( 1)  - [2.corrupttreat]1bn._at + [2.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0841315   .0449776     1.87   0.061    -.0040477    .1723107
------------------------------------------------------------------------------

 ( 1)  - [3.corrupttreat]1bn._at + [3.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0178566   .0414365     0.43   0.667    -.0633802    .0990934
------------------------------------------------------------------------------

 ( 1)  - [4.corrupttreat]1bn._at + [4.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0813212   .0494292     1.65   0.100    -.0155853    .1782277
------------------------------------------------------------------------------

. clear

. svmat ests
number of observations will be reset to 10
Press any key to continue, or Break to abort
number of observations (_N) was 0, now 10

. gen n = .
(10 missing values generated)

. for any 1.1 1.5 2 3.1 3.5 4 5.1 5.5 6 6.2 \ any 1 2 3 4 5 6 7 8 9 10: replace n = X in Y

->  replace n = 1.1 in 1
(1 real change made)

->  replace n = 1.5 in 2
(1 real change made)

->  replace n = 2 in 3
(1 real change made)

->  replace n = 3.1 in 4
(1 real change made)

->  replace n = 3.5 in 5
(1 real change made)

->  replace n = 4 in 6
(1 real change made)

->  replace n = 5.1 in 7
(1 real change made)

->  replace n = 5.5 in 8
(1 real change made)

->  replace n = 6 in 9
(1 real change made)

->  replace n = 6.2 in 10
(1 real change made)

. svmat diffs

. gen n2 = .
(10 missing values generated)

. for any 1.75 3.75 5.75 \ any 1 2 3: replace n2 = X in Y

->  replace n2 = 1.75 in 1
(1 real change made)

->  replace n2 = 3.75 in 2
(1 real change made)

->  replace n2 = 5.75 in 3
(1 real change made)

. tostring diffs1, gen(diffs_) force format(%9.3f)
diffs_ generated as str5
diffs_ was forced to string; some loss of information

. gen stars = "" in 1
(10 missing values generated)

. replace stars = "" in 2
(0 real changes made)

. replace stars = "*" in 3
(1 real change made)

. egen diffs = concat(diffs_ stars)

. gen t = "Diff.: " in 1/3
(7 missing values generated)

. egen lab = concat(t diffs)

. forval i = 1/3 {
  2.         loc l`i' = lab in `i'
  3.         di "`l`i''"
  4.         }
Diff.: 0.084
Diff.: 0.018
Diff.: 0.081*

. twoway (rcap ests2 ests3 n, msize(0) lcol(black) horizontal) ///
>                 (scatter n ests1, mcol(black) msize(medium) msymbol(O)), ///
>                 legend(off) scheme(plotplain) ///
>                 ylabel(1.1 "{bf:Bribes}" 1.5 "Not paid bribe" 2 "Paid bribe" ///
>                                 3.1 "{bf:Bribes common}" 3.5 "Not paid bribe" ///
>                                 4 "Paid bribe" 5.1 "{bf:Bribes but jobs}" ///
>                                 5.5 "Not paid bribe" 6 "Paid bribe", ///
>                                 labgap(6pt) noticks nogrid angle(0) labsize(small)) ytitle("") ///
>                 xscale(range(-0.45(.1).05)) xlabel(#6, glcolor(gs2)) xtitle("Effect on Pr(Voting for candidate)") ///
>                 yscale(reverse )  ytick(1.5 2 3.5 4 5.5 6, grid glcolor(gs2)) ///
>                 xline(0, lcolor(black) lpattern(dash) lwidth(vthin))  ///
>                 text(1.71 -.18 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(small) color(black)) ///
>                 text(1.75 -.13 "`l1'", size(small)) ///
>                 text(3.71 -.25 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(small) color(black)) ///
>                 text(3.75 -.20 "`l2'", size(small)) ///
>                 text(5.73 -.08 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(small) color(black)) ///
>                 text(5.75 -.03 "`l3'", size(small)) 

. 
. 
. *** Figure A5: Corruption Treatment Effects, by Information Source
. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. reg vote i.corrupttreat##i.copsource2 badeconomy##arias i.copartisan i.female i.country ///
>         [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(18, 4333)       =      55.60
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1110
                                                Root MSE          =     .42767

                                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
---------------------------------------------------------------------------------------------------------
                                        |               Robust
                                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------------------------+----------------------------------------------------------------
                           corrupttreat |
                                Bribes  |   -.366893    .023297   -15.75   0.000     -.412567   -.3212191
                         Bribes common  |  -.3870434   .0231386   -16.73   0.000     -.432407   -.3416799
                       Bribes but jobs  |  -.3008973   .0244268   -12.32   0.000    -.3487863   -.2530083
                                        |
                             copsource2 |
                Out-partisan newspaper  |  -.0297083   .0245503    -1.21   0.226    -.0778394    .0184229
                 Co-partisan newspaper  |   .0021127   .0301545     0.07   0.944    -.0570055     .061231
                                        |
                corrupttreat#copsource2 |
         Bribes#Out-partisan newspaper  |   .0300945   .0307359     0.98   0.328    -.0301637    .0903526
          Bribes#Co-partisan newspaper  |  -.0017825   .0379737    -0.05   0.963    -.0762304    .0726655
  Bribes common#Out-partisan newspaper  |   .0276735   .0304472     0.91   0.363    -.0320186    .0873655
   Bribes common#Co-partisan newspaper  |   .0551854   .0388438     1.42   0.155    -.0209683     .131339
Bribes but jobs#Out-partisan newspaper  |   .0539873   .0323931     1.67   0.096    -.0095198    .1174945
 Bribes but jobs#Co-partisan newspaper  |   .0543921   .0414195     1.31   0.189    -.0268114    .1355956
                                        |
                             badeconomy |
                      Economy worsened  |  -.0675262   .0130327    -5.18   0.000    -.0930769   -.0419755
                                1.arias |  -.0526582   .0153466    -3.43   0.001    -.0827454    -.022571
                                        |
                       badeconomy#arias |
                    Economy worsened#1  |   .0961794   .0214744     4.48   0.000     .0540785    .1382803
                                        |
                             copartisan |
                 Co-partisan candidate  |   .0766712   .0188566     4.07   0.000     .0397025    .1136399
                                        |
                                 female |
                                Female  |  -.0193341   .0092698    -2.09   0.037    -.0375077   -.0011605
                                        |
                                country |
                                 Chile  |  -.0081567   .0090169    -0.90   0.366    -.0258344    .0095209
                               Uruguay  |   .0022613   .0091271     0.25   0.804    -.0156325     .020155
                                        |
                                  _cons |   .5920945   .0211065    28.05   0.000      .550715     .633474
---------------------------------------------------------------------------------------------------------

. margins, dydx(corrupttreat) at(copsource2 = (0 1 2)) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat 3.corrupttreat 4.corrupttreat

1._at        : copsource2      =           0

2._at        : copsource2      =           1

3._at        : copsource2      =           2

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |   -.366893    .023297   -15.75   0.000     -.412567   -.3212191
             2  |  -.3367986   .0201165   -16.74   0.000    -.3762373   -.2973599
             3  |  -.3686755   .0292906   -12.59   0.000    -.4261001    -.311251
----------------+----------------------------------------------------------------
3.corrupttreat  |
            _at |
             1  |  -.3870434   .0231386   -16.73   0.000     -.432407   -.3416799
             2  |  -.3593699   .0197655   -18.18   0.000    -.3981205   -.3206194
             3  |  -.3318581   .0308662   -10.75   0.000    -.3923716   -.2713445
----------------+----------------------------------------------------------------
4.corrupttreat  |
            _at |
             1  |  -.3008973   .0244268   -12.32   0.000    -.3487863   -.2530083
             2  |    -.24691   .0211919   -11.65   0.000    -.2884569    -.205363
             3  |  -.2465052    .033383    -7.38   0.000    -.3119529   -.1810575
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. mat ests = J(13,3,.)

. local j 2 3 4 6 7 8 10 11 12

. local k 4 5 6 7 8 9 10 11 12

. forval i = 1/9 {
  2.         loc m : word `i' of `j'
  3.         loc n : word `i' of `k'
  4.         mat ests[`m',1] = tab[1,`n']
  5.         mat ests[`m',2] = tab[5,`n']
  6.         mat ests[`m',3] = tab[6,`n']
  7.         }

. clear

. svmat ests
number of observations will be reset to 13
Press any key to continue, or Break to abort
number of observations (_N) was 0, now 13

. gen n = .
(13 missing values generated)

. for any 1.1 1.5 2 2.5 3.6 4 4.5 5 6.1 6.5 7 7.5 7.7 \ any 1 2 3 4 5 6 7 8 9 10 11 12 13: replace n = X in Y

->  replace n = 1.1 in 1
(1 real change made)

->  replace n = 1.5 in 2
(1 real change made)

->  replace n = 2 in 3
(1 real change made)

->  replace n = 2.5 in 4
(1 real change made)

->  replace n = 3.6 in 5
(1 real change made)

->  replace n = 4 in 6
(1 real change made)

->  replace n = 4.5 in 7
(1 real change made)

->  replace n = 5 in 8
(1 real change made)

->  replace n = 6.1 in 9
(1 real change made)

->  replace n = 6.5 in 10
(1 real change made)

->  replace n = 7 in 11
(1 real change made)

->  replace n = 7.5 in 12
(1 real change made)

->  replace n = 7.7 in 13
(1 real change made)

. twoway (rcap ests2 ests3 n, msize(0) lcol(black) horizontal) ///
>                 (scatter n ests1, mcol(black) msize(medium) msymbol(O)), ///
>                 legend(off) scheme(plotplain) ///
>                 ylabel(1.1 "{bf:Bribes}" 1.5 "Judicial officials" 2 "Out-partisan newspaper" ///
>                                 2.5 "Co-partisan newspaper" 3.6 "{bf:Bribes common}" 4 "Judicial officials" ///
>                                 4.5 "Out-partisan newspaper" 5 "Co-partisan newspaper" 6.1 "{bf:Bribes but jobs}" ///
>                                 6.5 "Judicial officials" 7 "Out-partisan newspaper" 7.5 "Co-partisan newspaper", ///
>                                 labgap(6pt) noticks nogrid angle(0) labsize(small)) ytitle("") ///
>                 xscale(range(-0.5(.1)-0.1)) xlabel(#6, glcolor(gs2)) xtitle("Effect on Pr(Voting for candidate)") ///
>                 yscale(reverse )  ytick(1.5 2 2.5 4 4.5 5 6.5 7 7.5, grid glcolor(gs2)) 

. 
. 
. *** Figure A6: What Conditions Amplify the Mitigating Effect of Side Benefits?
. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. mat ests = J(13,3,.)

. mat diffs = J(6,1,.)

. * state of the economy
. reg vote i.corrupttreat##badeconomy##inc i.copartisan i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(21, 4333)       =      47.70
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1108
                                                Root MSE          =      .4278

                                                   (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------------------
                                    |               Robust
                               vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------------------+----------------------------------------------------------------
                       corrupttreat |
                            Bribes  |  -.3632061   .0266847   -13.61   0.000    -.4155217   -.3108905
                     Bribes common  |  -.3736196   .0264598   -14.12   0.000    -.4254944   -.3217448
                   Bribes but jobs  |  -.2535878   .0286488    -8.85   0.000    -.3097542   -.1974214
                                    |
                         badeconomy |
                  Economy worsened  |   .0324421   .0307427     1.06   0.291    -.0278294    .0927136
                                    |
            corrupttreat#badeconomy |
           Bribes#Economy worsened  |   .0136584   .0385152     0.35   0.723    -.0618512    .0891679
    Bribes common#Economy worsened  |  -.0054462   .0380869    -0.14   0.886    -.0801161    .0692236
  Bribes but jobs#Economy worsened  |  -.0211713    .040826    -0.52   0.604    -.1012111    .0588685
                                    |
                              1.inc |   .0571753   .0323982     1.76   0.078    -.0063419    .1206924
                                    |
                   corrupttreat#inc |
                          Bribes#1  |   .0072959   .0385306     0.19   0.850    -.0682439    .0828356
                   Bribes common#1  |    .013979   .0391137     0.36   0.721    -.0627038    .0906618
                 Bribes but jobs#1  |   -.038628   .0405473    -0.95   0.341    -.1181214    .0408654
                                    |
                     badeconomy#inc |
                Economy worsened#1  |  -.1204672    .045995    -2.62   0.009    -.2106409   -.0302934
                                    |
        corrupttreat#badeconomy#inc |
         Bribes#Economy worsened#1  |  -.0033333   .0544855    -0.06   0.951    -.1101528    .1034862
  Bribes common#Economy worsened#1  |   .0245072   .0545707     0.45   0.653    -.0824794    .1314938
Bribes but jobs#Economy worsened#1  |   .0741162   .0577153     1.28   0.199    -.0390354    .1872678
                                    |
                         copartisan |
             Co-partisan candidate  |    .078128   .0188517     4.14   0.000      .041169    .1150869
                                    |
                         copsource2 |
            Out-partisan newspaper  |  -.0015406   .0103123    -0.15   0.881     -.021758    .0186767
             Co-partisan newspaper  |   .0282591   .0130869     2.16   0.031     .0026021    .0539162
                                    |
                             female |
                            Female  |   -.019114   .0092651    -2.06   0.039    -.0372783   -.0009497
                                    |
                            country |
                             Chile  |  -.0084414   .0090233    -0.94   0.350    -.0261317     .009249
                           Uruguay  |   .0020303   .0091524     0.22   0.824    -.0159131    .0199737
                                    |
                              _cons |   .5229993   .0234417    22.31   0.000     .4770416     .568957
-----------------------------------------------------------------------------------------------------

. margins, dydx(2.corrupttreat 4.corrupttreat) at(inc = 1 badeconomy = (0 1)) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat 4.corrupttreat

1._at        : badeconomy      =           0
               inc             =           1

2._at        : badeconomy      =           1
               inc             =           1

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |  -.3559102   .0273574   -13.01   0.000    -.4095447   -.3022757
             2  |  -.3455851    .026239   -13.17   0.000    -.3970271   -.2941432
----------------+----------------------------------------------------------------
4.corrupttreat  |
            _at |
             1  |  -.2922158   .0290255   -10.07   0.000    -.3491206   -.2353111
             2  |  -.2392709   .0289495    -8.27   0.000    -.2960268    -.182515
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. local j = 3

. foreach i in 2 3 5 6 {
  2.         mat ests[`i',1] = tab[1,`j']
  3.         mat ests[`i',2] = tab[5,`j']
  4.         mat ests[`i',3] = tab[6,`j']
  5.         local j = `j' + 1
  6.         }

. lincom _b[2.corrupttreat:2._at]-_b[2.corrupttreat:1._at]

 ( 1)  - [2.corrupttreat]1bn._at + [2.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0103251   .0378967     0.27   0.785    -.0639719     .084622
------------------------------------------------------------------------------

. mat diffs[1,1] = r(estimate)

. lincom _b[4.corrupttreat:2._at]-_b[4.corrupttreat:1._at]

 ( 1)  - [4.corrupttreat]1bn._at + [4.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0529449   .0409813     1.29   0.196    -.0273993    .1332891
------------------------------------------------------------------------------

. mat diffs[2,1] = r(estimate)

. lincom _b[4.corrupttreat:2._at]-_b[4.corrupttreat:1._at]-(_b[2.corrupttreat:2._at]-_b[2.corrupttreat:1._at])

 ( 1)  [2.corrupttreat]1bn._at - [2.corrupttreat]2._at - [4.corrupttreat]1bn._at + [4.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0426198   .0356111     1.20   0.231    -.0271961    .1124358
------------------------------------------------------------------------------

. mat diffs[3,1] = r(estimate)

. * employment status
. reg vote i.corrupttreat##i.unemp badeconomy##arias i.copartisan i.copsource2 i.female i.country [pw=weight1500], cl(uniq_id)
(sum of wgt is 5,205.24843931198)

Linear regression                               Number of obs     =      5,398
                                                F(16, 2698)       =      40.04
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1128
                                                Root MSE          =     .43013

                                       (Std. Err. adjusted for 2,699 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3418675   .0183559   -18.62   0.000    -.3778606   -.3058744
         Bribes common  |  -.3809183   .0178168   -21.38   0.000    -.4158542   -.3459824
       Bribes but jobs  |  -.2725858   .0198123   -13.76   0.000    -.3114347   -.2337369
                        |
                1.unemp |   .0095188   .0380117     0.25   0.802    -.0650162    .0840538
                        |
     corrupttreat#unemp |
              Bribes#1  |  -.0326631   .0506264    -0.65   0.519    -.1319335    .0666073
       Bribes common#1  |   .0265789   .0507167     0.52   0.600    -.0728687    .1260264
     Bribes but jobs#1  |   .0273328   .0547939     0.50   0.618    -.0801095    .1347751
                        |
             badeconomy |
      Economy worsened  |  -.0595808   .0166117    -3.59   0.000    -.0921538   -.0270078
                1.arias |   -.041719   .0194054    -2.15   0.032      -.07977   -.0036681
                        |
       badeconomy#arias |
    Economy worsened#1  |   .0863785   .0274313     3.15   0.002       .03259     .140167
                        |
             copartisan |
 Co-partisan candidate  |   .0800809   .0251286     3.19   0.001     .0308076    .1293541
                        |
             copsource2 |
Out-partisan newspaper  |   .0014648    .013156     0.11   0.911    -.0243321    .0272617
 Co-partisan newspaper  |   .0254911   .0166086     1.53   0.125    -.0070757    .0580579
                        |
                 female |
                Female  |  -.0172193   .0117915    -1.46   0.144    -.0403406     .005902
                        |
                country |
                 Chile  |  -.0172869   .0115982    -1.49   0.136    -.0400291    .0054553
               Uruguay  |   .0072679   .0115003     0.63   0.527    -.0152824    .0298182
                        |
                  _cons |   .5726044   .0208089    27.52   0.000     .5318015    .6134073
-----------------------------------------------------------------------------------------

. margins, dydx(2.corrupttreat 4.corrupttreat) at(unemp = (0 1)) post

Average marginal effects                        Number of obs     =      5,398
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat 4.corrupttreat

1._at        : unemp           =           0

2._at        : unemp           =           1

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |  -.3418675   .0183559   -18.62   0.000    -.3778606   -.3058744
             2  |  -.3745306   .0471538    -7.94   0.000    -.4669918   -.2820693
----------------+----------------------------------------------------------------
4.corrupttreat  |
            _at |
             1  |  -.2725858   .0198123   -13.76   0.000    -.3114347   -.2337369
             2  |   -.245253   .0510637    -4.80   0.000    -.3453808   -.1451251
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. local j = 3

. foreach i in 8 9 11 12 {
  2.         mat ests[`i',1] = tab[1,`j']
  3.         mat ests[`i',2] = tab[5,`j']
  4.         mat ests[`i',3] = tab[6,`j']
  5.         local j = `j' + 1
  6.         }

. lincom _b[2.corrupttreat:2._at]-_b[2.corrupttreat:1._at]

 ( 1)  - [2.corrupttreat]1bn._at + [2.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |  -.0326631   .0506264    -0.65   0.519    -.1319335    .0666073
------------------------------------------------------------------------------

. mat diffs[4,1] = r(estimate)

. lincom _b[4.corrupttreat:2._at]-_b[4.corrupttreat:1._at]

 ( 1)  - [4.corrupttreat]1bn._at + [4.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0273328   .0547939     0.50   0.618    -.0801095    .1347751
------------------------------------------------------------------------------

. mat diffs[5,1] = r(estimate)

. lincom _b[4.corrupttreat:2._at]-_b[4.corrupttreat:1._at]-(_b[2.corrupttreat:2._at]-_b[2.corrupttreat:1._at])

 ( 1)  [2.corrupttreat]1bn._at - [2.corrupttreat]2._at - [4.corrupttreat]1bn._at + [4.corrupttreat]2._at = 0

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0599959   .0484505     1.24   0.216    -.0350079    .1549997
------------------------------------------------------------------------------

. mat diffs[6,1] = r(estimate)

. * creating the graph
. clear

. svmat ests
number of observations will be reset to 13
Press any key to continue, or Break to abort
number of observations (_N) was 0, now 13

. gen n = .
(13 missing values generated)

. for any 1.1 1.5 2 3.1 3.5 4 5.6 6 6.5 7.6 8 8.5 8.7 \ any 1 2 3 4 5 6 7 8 9 10 11 12 13: replace n = X in Y

->  replace n = 1.1 in 1
(1 real change made)

->  replace n = 1.5 in 2
(1 real change made)

->  replace n = 2 in 3
(1 real change made)

->  replace n = 3.1 in 4
(1 real change made)

->  replace n = 3.5 in 5
(1 real change made)

->  replace n = 4 in 6
(1 real change made)

->  replace n = 5.6 in 7
(1 real change made)

->  replace n = 6 in 8
(1 real change made)

->  replace n = 6.5 in 9
(1 real change made)

->  replace n = 7.6 in 10
(1 real change made)

->  replace n = 8 in 11
(1 real change made)

->  replace n = 8.5 in 12
(1 real change made)

->  replace n = 8.7 in 13
(1 real change made)

. svmat diffs

. gen n2 = .
(13 missing values generated)

. for any 1.75 3.75 2.75 6.25 8.25 7.25 \ any 1 2 3 4 5 6: replace n2 = X in Y

->  replace n2 = 1.75 in 1
(1 real change made)

->  replace n2 = 3.75 in 2
(1 real change made)

->  replace n2 = 2.75 in 3
(1 real change made)

->  replace n2 = 6.25 in 4
(1 real change made)

->  replace n2 = 8.25 in 5
(1 real change made)

->  replace n2 = 7.25 in 6
(1 real change made)

. tostring diffs1, gen(diffs_) force format(%9.3f)
diffs_ generated as str6
diffs_ was forced to string; some loss of information

. gen stars = "" in 1/6
(13 missing values generated)

. egen diffs = concat(diffs_ stars)

. gen t = "Diff.: " in 1/6
(7 missing values generated)

. egen lab = concat(t diffs)

. forval i = 1/6 {
  2.         loc l`i' = lab in `i'
  3.         di "`l`i''"
  4.         }
Diff.: 0.010
Diff.: 0.053
Diff.: 0.043
Diff.: -0.033
Diff.: 0.027
Diff.: 0.060

. twoway (rcap ests2 ests3 n, msize(0) lcol(black) horizontal) ///
>                 (scatter n ests1, mcol(black) msize(medium) msymbol(O)), ///
>                 legend(off) scheme(plotplain) ///
>                 ylabel(1.1 "{bf: Improving economy}" 1.5 "Bribes" 2 "Bribes but jobs" ///
>                                 3.1 "{bf: Worsening economy}" 3.5 "Bribes" 4 "Bribes but jobs" ///
>                                 5.6 "{bf: Employed}" 6 "Bribes" 6.5 "Bribes but jobs" ///
>                                 7.6 "{bf: Unemployed}" 8 "Bribes" 8.5 "Bribes but jobs", ///
>                                 labgap(5pt) noticks nogrid angle(0) labsize(small)) ytitle("") ///
>                 xscale(range(-0.6(.1).1)) xlabel(#6, glcolor(gs2)) xtitle("Effect on Pr(Voting for candidate)") ///
>                 yscale(reverse )  ytick(1.5 2 3.5 4 6 6.5 8 8.5, grid glcolor(gs2)) ///
>                 xline(0, lcolor(black) lpattern(dash) lwidth(vthin)) ///
>                 text(1.73 -.27 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(vsmall) color(black)) ///
>                 text(1.75 -.20 "`l1'", size(small)) ///
>                 text(3.73 -.17 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(vsmall) color(black)) ///
>                 text(3.75 -.10 "`l2'", size(small)) ///
>                 text(2.63 .025 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(huge) color(black)) ///
>                 text(2.79 .12 "`l3'", size(medium)) ///
>                 text(6.23 -.255 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(vsmall) color(black)) ///
>                 text(6.25 -.185 "`l4'", size(small)) ///
>                 text(8.23 -.125 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(vsmall) color(black)) ///
>                 text(8.25 -.055 "`l5'", size(small)) ///
>                 text(7.13 .025 "`=ustrunescape("\u23AB")'" "`=ustrunescape("\u23AC")'" "`=ustrunescape("\u23AD")'" , size(huge) color(black)) ///
>                 text(7.25 .12 "`l6'", size(medium)) ///
>                 yline(5, lpattern(solid) lcolor(gray) lwidth(vthin))

. 
. 
. *** Figure A7: Economy Treatment Effect, by Country and Respondents’ Retrospective Economic Perceptions
. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. mat ests = J(19,3,.)

. * overall, pooled
. reg vote i.corrupttreat badeconomy##arias i.copartisan i.copsource2 i.female i.country ///
>         [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(12, 4333)       =      82.21
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1103
                                                Root MSE          =      .4277

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3533974    .013401   -26.37   0.000    -.3796702   -.3271247
         Bribes common  |  -.3631275   .0134214   -27.06   0.000    -.3894404   -.3368146
       Bribes but jobs  |   -.264988   .0145094   -18.26   0.000    -.2934338   -.2365422
                        |
             badeconomy |
      Economy worsened  |  -.0676445   .0130233    -5.19   0.000    -.0931768   -.0421122
                1.arias |  -.0528361   .0153538    -3.44   0.001    -.0829373   -.0227349
                        |
       badeconomy#arias |
    Economy worsened#1  |   .0968282    .021476     4.51   0.000     .0547242    .1389322
                        |
             copartisan |
 Co-partisan candidate  |   .0775461   .0188731     4.11   0.000     .0405451     .114547
                        |
             copsource2 |
Out-partisan newspaper  |  -.0015192   .0103112    -0.15   0.883    -.0217345    .0186961
 Co-partisan newspaper  |   .0283579   .0130783     2.17   0.030     .0027178    .0539981
                        |
                 female |
                Female  |  -.0191172   .0092638    -2.06   0.039    -.0372791   -.0009553
                        |
                country |
                 Chile  |  -.0083276   .0090087    -0.92   0.355    -.0259894    .0093341
               Uruguay  |   .0021432   .0091276     0.23   0.814    -.0157516    .0200379
                        |
                  _cons |   .5735772   .0160463    35.75   0.000     .5421182    .6050361
-----------------------------------------------------------------------------------------

. margins, dydx(badeconomy) at(arias == 0) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 1.badeconomy
at           : arias           =           0

-----------------------------------------------------------------------------------
                  |            Delta-method
                  |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------+----------------------------------------------------------------
       badeconomy |
Economy worsened  |  -.0676445   .0130233    -5.19   0.000    -.0931768   -.0421122
-----------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. mat ests[2,1] = tab[1,2]

. mat ests[2,2] = tab[5,2]

. mat ests[2,3] = tab[6,2]

. * overall, by country
. reg vote i.corrupttreat badeconomy##arias##i.country i.copartisan i.copsource2 i.female ///
>         [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(18, 4333)       =      55.83
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1117
                                                Root MSE          =      .4275

                                           (Std. Err. adjusted for 4,334 clusters in uniq_id)
---------------------------------------------------------------------------------------------
                            |               Robust
                       vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------------+----------------------------------------------------------------
               corrupttreat |
                    Bribes  |   -.352898   .0134054   -26.33   0.000    -.3791794   -.3266166
             Bribes common  |   -.362222   .0134263   -26.98   0.000    -.3885444   -.3358997
           Bribes but jobs  |  -.2645978   .0145009   -18.25   0.000    -.2930269   -.2361687
                            |
                 badeconomy |
          Economy worsened  |  -.0613259   .0225573    -2.72   0.007    -.1055498    -.017102
                    1.arias |  -.0021163   .0276983    -0.08   0.939    -.0564191    .0521865
                            |
           badeconomy#arias |
        Economy worsened#1  |   .0657024   .0384093     1.71   0.087    -.0095994    .1410042
                            |
                    country |
                     Chile  |   .0075418   .0231222     0.33   0.744    -.0377895    .0528731
                   Uruguay  |   .0478338   .0234523     2.04   0.041     .0018554    .0938123
                            |
         badeconomy#country |
    Economy worsened#Chile  |   .0016284   .0317177     0.05   0.959    -.0605544    .0638113
  Economy worsened#Uruguay  |  -.0200264   .0321139    -0.62   0.533    -.0829861    .0429333
                            |
              arias#country |
                   1#Chile  |  -.0412382    .037877    -1.09   0.276    -.1154965    .0330201
                 1#Uruguay  |   -.108655    .038149    -2.85   0.004    -.1834466   -.0338634
                            |
   badeconomy#arias#country |
  Economy worsened#1#Chile  |   .0160034   .0526283     0.30   0.761     -.087175    .1191819
Economy worsened#1#Uruguay  |   .0757064   .0535051     1.41   0.157    -.0291909    .1806037
                            |
                 copartisan |
     Co-partisan candidate  |   .0763023   .0188531     4.05   0.000     .0393406     .113264
                            |
                 copsource2 |
    Out-partisan newspaper  |  -.0018649   .0103202    -0.18   0.857    -.0220978    .0183681
     Co-partisan newspaper  |   .0283638   .0130681     2.17   0.030     .0027437    .0539839
                            |
                     female |
                    Female  |  -.0189334   .0092628    -2.04   0.041    -.0370933   -.0007736
                      _cons |    .552438   .0203719    27.12   0.000     .5124986    .5923774
---------------------------------------------------------------------------------------------

. margins, dydx(badeconomy) at(arias == 0 country = (1 2 3)) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 1.badeconomy

1._at        : arias           =           0
               country         =           1

2._at        : arias           =           0
               country         =           2

3._at        : arias           =           0
               country         =           3

-------------------------------------------------------------------------------
              |            Delta-method
              |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
--------------+----------------------------------------------------------------
0.badeconomy  |  (base outcome)
--------------+----------------------------------------------------------------
1.badeconomy  |
          _at |
           1  |  -.0613259   .0225573    -2.72   0.007    -.1055498    -.017102
           2  |  -.0596975   .0222754    -2.68   0.007    -.1033687   -.0160262
           3  |  -.0813523   .0228394    -3.56   0.000    -.1261293   -.0365753
-------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. forval i = 3/5 {
  2.         loc j = `i'+1
  3.         mat ests[`i',1] = tab[1,`j']
  4.         mat ests[`i',2] = tab[5,`j']
  5.         mat ests[`i',3] = tab[6,`j']
  6.         }

. * by economic perceptions, pooled
. reg vote i.corrupttreat badeconomy##arias##i.econperc i.copartisan i.copsource2 i.female i.country ///
>         [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,239.65765440464)

Linear regression                               Number of obs     =      8,548
                                                F(16, 4273)       =      60.97
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1102
                                                Root MSE          =     .42828

                                         (Std. Err. adjusted for 4,274 clusters in uniq_id)
-------------------------------------------------------------------------------------------
                          |               Robust
                     vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
--------------------------+----------------------------------------------------------------
             corrupttreat |
                  Bribes  |  -.3526529   .0135011   -26.12   0.000     -.379122   -.3261839
           Bribes common  |  -.3630878   .0135125   -26.87   0.000    -.3895793   -.3365962
         Bribes but jobs  |  -.2645863   .0146137   -18.11   0.000    -.2932367    -.235936
                          |
               badeconomy |
        Economy worsened  |  -.0779347    .018758    -4.15   0.000      -.11471   -.0411593
                  1.arias |  -.0685675   .0219105    -3.13   0.002    -.1115234   -.0256115
                          |
         badeconomy#arias |
      Economy worsened#1  |   .1275304   .0310782     4.10   0.000      .066601    .1884597
                          |
               1.econperc |  -.0164915   .0191585    -0.86   0.389     -.054052    .0210691
                          |
      badeconomy#econperc |
      Economy worsened#1  |   .0202639   .0262675     0.77   0.440    -.0312341    .0717619
                          |
           arias#econperc |
                     1 1  |   .0302033   .0309925     0.97   0.330     -.030558    .0909647
                          |
badeconomy#arias#econperc |
    Economy worsened#1#1  |  -.0562726   .0433951    -1.30   0.195    -.1413496    .0288044
                          |
               copartisan |
   Co-partisan candidate  |    .079245   .0190674     4.16   0.000     .0418629     .116627
                          |
               copsource2 |
  Out-partisan newspaper  |  -.0013682   .0104202    -0.13   0.896    -.0217973    .0190608
   Co-partisan newspaper  |   .0263583   .0131622     2.00   0.045     .0005535     .052163
                          |
                   female |
                  Female  |  -.0174008   .0093472    -1.86   0.063    -.0357262    .0009246
                          |
                  country |
                   Chile  |  -.0113066    .009173    -1.23   0.218    -.0292904    .0066771
                 Uruguay  |   .0000965   .0092427     0.01   0.992     -.018024    .0182169
                          |
                    _cons |    .583276   .0190612    30.60   0.000     .5459062    .6206458
-------------------------------------------------------------------------------------------

. margins, dydx(badeconomy) at(arias == 0 econperc = (0 1)) post

Average marginal effects                        Number of obs     =      8,548
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 1.badeconomy

1._at        : arias           =           0
               econperc        =           0

2._at        : arias           =           0
               econperc        =           1

-------------------------------------------------------------------------------
              |            Delta-method
              |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
--------------+----------------------------------------------------------------
0.badeconomy  |  (base outcome)
--------------+----------------------------------------------------------------
1.badeconomy  |
          _at |
           1  |  -.0779347    .018758    -4.15   0.000      -.11471   -.0411593
           2  |  -.0576707    .018375    -3.14   0.002    -.0936953   -.0216461
-------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. forval i = 8/9 {
  2.         loc j = `i'-5
  3.         mat ests[`i',1] = tab[1,`j']
  4.         mat ests[`i',2] = tab[5,`j']
  5.         mat ests[`i',3] = tab[6,`j']
  6.         }

. * by economic perceptions, by country
. reg vote i.corrupttreat i.country##badeconomy##arias##i.econperc i.copartisan i.copsource2 i.female ///
>         [pw=weight1500], cl(uniq_id)
(sum of wgt is 8,239.65765440464)

Linear regression                               Number of obs     =      8,548
                                                F(30, 4273)       =      34.70
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1133
                                                Root MSE          =     .42788

                                                 (Std. Err. adjusted for 4,274 clusters in uniq_id)
---------------------------------------------------------------------------------------------------
                                  |               Robust
                             vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------------------+----------------------------------------------------------------
                     corrupttreat |
                          Bribes  |  -.3522593   .0134805   -26.13   0.000    -.3786881   -.3258305
                   Bribes common  |  -.3629722   .0135059   -26.88   0.000    -.3894508   -.3364936
                 Bribes but jobs  |  -.2647409   .0145938   -18.14   0.000    -.2933523   -.2361294
                                  |
                          country |
                           Chile  |   .0293635   .0343456     0.85   0.393    -.0379718    .0966988
                         Uruguay  |   .0763094   .0350141     2.18   0.029     .0076636    .1449552
                                  |
                       badeconomy |
                Economy worsened  |  -.1011163   .0347711    -2.91   0.004    -.1692857   -.0329469
                                  |
               country#badeconomy |
          Chile#Economy worsened  |   .0431511   .0463596     0.93   0.352    -.0477378    .1340399
        Uruguay#Economy worsened  |   .0154861   .0470126     0.33   0.742     -.076683    .1076552
                                  |
                          1.arias |   .0308023   .0445501     0.69   0.489     -.056539    .1181436
                                  |
                    country#arias |
                         Chile#1  |  -.0734708   .0569638    -1.29   0.197    -.1851495    .0382079
                       Uruguay#1  |  -.1919288   .0567787    -3.38   0.001    -.3032446    -.080613
                                  |
                 badeconomy#arias |
              Economy worsened#1  |    .078463   .0611491     1.28   0.200     -.041421     .198347
                                  |
         country#badeconomy#arias |
        Chile#Economy worsened#1  |   .0081534   .0790732     0.10   0.918    -.1468711     .163178
      Uruguay#Economy worsened#1  |    .125606   .0798942     1.57   0.116    -.0310282    .2822402
                                  |
                       1.econperc |   .0251341   .0339241     0.74   0.459    -.0413748     .091643
                                  |
                 country#econperc |
                         Chile#1  |  -.0459301    .047315    -0.97   0.332     -.138692    .0468319
                       Uruguay#1  |  -.0644268   .0475611    -1.35   0.176    -.1576713    .0288176
                                  |
              badeconomy#econperc |
              Economy worsened#1  |   .0562951   .0456508     1.23   0.218    -.0332042    .1457945
                                  |
      country#badeconomy#econperc |
        Chile#Economy worsened#1  |  -.0634963   .0641032    -0.99   0.322     -.189172    .0621793
      Uruguay#Economy worsened#1  |  -.0372348   .0647241    -0.58   0.565    -.1641277    .0896582
                                  |
                   arias#econperc |
                             1 1  |   -.058177   .0571842    -1.02   0.309    -.1702878    .0539338
                                  |
           country#arias#econperc |
                       Chile#1#1  |   .0529901   .0773127     0.69   0.493     -.098583    .2045632
                     Uruguay#1#1  |   .1754366   .0780256     2.25   0.025      .022466    .3284072
                                  |
        badeconomy#arias#econperc |
            Economy worsened#1#1  |  -.0100547   .0789428    -0.13   0.899    -.1648236    .1447143
                                  |
country#badeconomy#arias#econperc |
      Chile#Economy worsened#1#1  |   .0114492   .1070956     0.11   0.915    -.1985138    .2214122
    Uruguay#Economy worsened#1#1  |  -.1353601   .1089712    -1.24   0.214    -.3490003    .0782802
                                  |
                       copartisan |
           Co-partisan candidate  |   .0784682   .0190469     4.12   0.000     .0411264      .11581
                                  |
                       copsource2 |
          Out-partisan newspaper  |  -.0018812   .0104267    -0.18   0.857    -.0223231    .0185606
           Co-partisan newspaper  |   .0260508   .0131409     1.98   0.047     .0002878    .0518139
                                  |
                           female |
                          Female  |  -.0175738   .0093377    -1.88   0.060    -.0358806     .000733
                            _cons |   .5398162    .029346    18.39   0.000     .4822829    .5973495
---------------------------------------------------------------------------------------------------

. margins, dydx(badeconomy) at(arias == 0 econperc = (0 1) country = (1 2 3)) post

Average marginal effects                        Number of obs     =      8,548
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 1.badeconomy

1._at        : country         =           1
               arias           =           0
               econperc        =           0

2._at        : country         =           1
               arias           =           0
               econperc        =           1

3._at        : country         =           2
               arias           =           0
               econperc        =           0

4._at        : country         =           2
               arias           =           0
               econperc        =           1

5._at        : country         =           3
               arias           =           0
               econperc        =           0

6._at        : country         =           3
               arias           =           0
               econperc        =           1

-------------------------------------------------------------------------------
              |            Delta-method
              |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
--------------+----------------------------------------------------------------
0.badeconomy  |  (base outcome)
--------------+----------------------------------------------------------------
1.badeconomy  |
          _at |
           1  |  -.1011163   .0347711    -2.91   0.004    -.1692857   -.0329469
           2  |  -.0448212   .0295665    -1.52   0.130    -.1027869    .0131446
           3  |  -.0579652   .0306684    -1.89   0.059    -.1180912    .0021607
           4  |  -.0651664   .0329724    -1.98   0.048    -.1298094   -.0005234
           5  |  -.0856302   .0316275    -2.71   0.007    -.1476365   -.0236239
           6  |  -.0665699   .0333627    -2.00   0.046    -.1319781   -.0011616
-------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. forval i = 11/12 {
  2.         loc j = `i'-4
  3.         mat ests[`i',1] = tab[1,`j']
  4.         mat ests[`i',2] = tab[5,`j']
  5.         mat ests[`i',3] = tab[6,`j']
  6.         }

. forval i = 14/15 {
  2.         loc j = `i'-5
  3.         mat ests[`i',1] = tab[1,`j']
  4.         mat ests[`i',2] = tab[5,`j']
  5.         mat ests[`i',3] = tab[6,`j']
  6.         }

. forval i = 17/18 {
  2.         loc j = `i'-6
  3.         mat ests[`i',1] = tab[1,`j']
  4.         mat ests[`i',2] = tab[5,`j']
  5.         mat ests[`i',3] = tab[6,`j']
  6.         }

. clear

. svmat ests
number of observations will be reset to 19
Press any key to continue, or Break to abort
number of observations (_N) was 0, now 19

. gen n = _n

. replace n = n+1 if n > 5
(14 real changes made)

. twoway (rcap ests2 ests3 n, msize(0) lcol(black) horizontal) ///
>                 (scatter n ests1, mcol(black) msize(medium) msymbol(O)), ///
>                 legend(off) scheme(plotplain) ///
>                 ylabel(1 "{bf:Overall}" 2 "Pooled" 3 "Argentina" ///
>                         4 "Chile" 5 "Uruguay" 7 "{bf:By economic perceptions}" ///
>                         8 "{bf:Pooled}" 9 "Economy not worse" 10 "Economy worse" ///
>                         11 "{bf:Argentina}" 12 "Economy not worse" 13 "Economy worse" ///
>                         14 "{bf:Chile}" 15 "Economy not worse" 16 "Economy worse" ///
>                         17 "{bf:Uruguay}" 18 "Economy not worse" 19 "Economy worse", ///
>                                 labgap(6pt) noticks nogrid angle(0) labsize(small)) ytitle("") ///
>                 xscale(range(-0.2(.05).05)) xlabel(#6, glcolor(gs2)) xtitle("Effect on Pr(Voting for candidate)") ///
>                 yscale(reverse )  ytick(2 3 4 5 9 10 12 13 15 16 18 19, grid glcolor(gs2)) ///
>                 xline(0, lcolor(black) lpattern(solid) lwidth(vthin)) ///
>                 yline(6.5, lpattern(solid) lcolor(gray) lwidth(vthin))

. 
. 
. *** Figure A8: Conventional Analysis vs. ANOVA-based Approach by Egami and Imai (2019)
. * prep data for R to run the estimation
. * notes: analysis without weights; estimation doesn't converge with economy and country dummies included
. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. keep uniq_id vote corrupttreat copartisan copsource female 

. recode corrupttreat (4 = 1) (3 = 2) (2 = 3) (1 = 4), gen(bribe)
(9328 differences between corrupttreat and bribe)

. recode copartisan (0 = 1) (1 = 0)
(copartisan: 9334 changes made)

. lab def bribe 4 "NoBribes" 3 "Bribes" 2 "BribesCommon" 1 "BribesJobs", replace

. lab val bribe bribe

. lab def copartisan 1 "OutCand" 0 "CopCand", replace

. lab val copartisan copartisan

. lab def copsource 0 "Court" 1 "OutNews" 2 "CopNews", replace

. lab val copsource copsource

. lab def female 0 "Male" 1 "Female", replace

. lab val female female

. drop if vote == .
(666 observations deleted)

. rename uniq id

. drop corruptt

. order id vote bribe copartisan copsource female

. saveold temp-R, replace v(12)
(saving in Stata 12 format, which can be read by Stata 11 or 12)
file temp-R.dta saved

. 
. * now run 'FigureA8.R', which exports a csv file 'ei.csv' to be used for graphing
. 
. reg vote b1.copartisan##b4.bribe b1.copsource b1.female, cl(id)

Linear regression                               Number of obs     =      8,668
                                                F(10, 4333)       =      93.52
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1067
                                                Root MSE          =     .42838

                                          (Std. Err. adjusted for 4,334 clusters in id)
---------------------------------------------------------------------------------------
                      |               Robust
                 vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
           copartisan |
             CopCand  |   .1552232   .0415922     3.73   0.000     .0736812    .2367651
                      |
                bribe |
          BribesJobs  |  -.2540348   .0150137   -16.92   0.000    -.2834693   -.2246004
        BribesCommon  |  -.3574055   .0139629   -25.60   0.000    -.3847799   -.3300312
              Bribes  |  -.3438159   .0139219   -24.70   0.000    -.3711098   -.3165219
                      |
     copartisan#bribe |
  CopCand#BribesJobs  |  -.1309034   .0599583    -2.18   0.029    -.2484522   -.0133545
CopCand#BribesCommon  |  -.0717292   .0554439    -1.29   0.196    -.1804276    .0369691
      CopCand#Bribes  |  -.1103613   .0556804    -1.98   0.048    -.2195234   -.0011992
                      |
           copsource2 |
               Court  |   .0001339   .0103311     0.01   0.990    -.0201203    .0203881
             CopNews  |   .0297064   .0122209     2.43   0.015     .0057471    .0536657
                      |
               female |
                Male  |   .0185535   .0092754     2.00   0.046     .0003689    .0367381
                      |
                _cons |   .5088904   .0127896    39.79   0.000     .4838163    .5339645
---------------------------------------------------------------------------------------

. margins, dydx(bribe) at(copartisan = (0 1)) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 1.bribe 2.bribe 3.bribe

1._at        : copartisan      =           0

2._at        : copartisan      =           1

------------------------------------------------------------------------------
             |            Delta-method
             |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
1.bribe      |
         _at |
          1  |  -.3849382   .0579729    -6.64   0.000    -.4985948   -.2712816
          2  |  -.2540348   .0150137   -16.92   0.000    -.2834693   -.2246004
-------------+----------------------------------------------------------------
2.bribe      |
         _at |
          1  |  -.4291348   .0533578    -8.04   0.000    -.5337433   -.3245262
          2  |  -.3574055   .0139629   -25.60   0.000    -.3847799   -.3300312
-------------+----------------------------------------------------------------
3.bribe      |
         _at |
          1  |  -.4541772   .0536925    -8.46   0.000    -.5594418   -.3489125
          2  |  -.3438159   .0139219   -24.70   0.000    -.3711098   -.3165219
-------------+----------------------------------------------------------------
4.bribe      |  (base outcome)
------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. mat ests = J(9,3,.)

. local j 1 3 5

. forval i = 1/3 {
  2.         loc m : word `i' of `j'
  3.         loc p = `i' + 1
  4.         mat ests[`p',1] = tab[1,`m']
  5.         mat ests[`p',2] = tab[5,`m']
  6.         mat ests[`p',3] = tab[6,`m']
  7.         }

. local j 2 4 6

. forval i = 1/3 {
  2.         loc m : word `i' of `j'
  3.         loc p = `i' + 5
  4.         mat ests[`p',1] = tab[1,`m']
  5.         mat ests[`p',2] = tab[5,`m']
  6.         mat ests[`p',3] = tab[6,`m']
  7.         }

. clear

. * import file created by the R script
. import delimited ei.csv, varnames(1) clear
(13 vars, 4 obs)

. rename conditionaleffectscopartisancopc ests1_c1

. rename v4 ests2_c1

. rename v5 ests3_c1

. rename conditionaleffectscopartisanoutc ests1_c0

. rename v8 ests2_c0

. rename v9 ests3_c0

. keep ests*

. drop in 4
(1 observation deleted)

. mkmat ests1_c1-ests3_c1, mat(c1)

. mkmat ests1_c0-ests3_c0, mat(c0)

. mat c = J(1,3,.)

. mat ests_ei = c \ c1 \ c \ c0 \ c

. clear

. svmat ests
number of observations will be reset to 9
Press any key to continue, or Break to abort
number of observations (_N) was 0, now 9

. svmat ests_ei

. gen n = .
(9 missing values generated)

. for any 1 4 3 2 6 9 8 7 10 \ any 1 2 3 4 5 6 7 8 9: replace n = X in Y

->  replace n = 1 in 1
(1 real change made)

->  replace n = 4 in 2
(1 real change made)

->  replace n = 3 in 3
(1 real change made)

->  replace n = 2 in 4
(1 real change made)

->  replace n = 6 in 5
(1 real change made)

->  replace n = 9 in 6
(1 real change made)

->  replace n = 8 in 7
(1 real change made)

->  replace n = 7 in 8
(1 real change made)

->  replace n = 10 in 9
(1 real change made)

. twoway (rcap ests2 ests3 n, msize(0) lcol(black) horizontal) ///
>                 (scatter n ests1, mcol(black) msize(medium) msymbol(O)), ///
>                 legend(off) scheme(plotplain) ///
>                 ylabel(1 "{bf:Co-partisan candidate}" 2 "Bribes" 3 "Bribes common" ///
>                         4 "Bribes but jobs" 6 "{bf:Out-partisan candidate}" ///
>                         7 "Bribes" 8 "Bribes common" 9 "Bribes but jobs", ///
>                         labgap(6pt) noticks nogrid angle(0) labsize(small)) ytitle("") ///
>                 xscale(range(-0.6(.2)0)) xlabel(#4, glcolor(gs2)) xtitle("Effect on Pr(Voting for candidate)") ///
>                 yscale(reverse )  ytick(2 3 4 7 8 9, grid glcolor(gs2)) title(Conventional) ///
>                 xline(0, lcolor(black) lpattern(dash) lwidth(vthin)) name(g1, replace)

. twoway (rcap ests_ei2 ests_ei3 n, msize(0) lcol(gray) horizontal) ///
>                 (scatter n ests_ei1, mcol(gray) msize(medium) msymbol(T)), ///
>                 legend(off) scheme(plotplain) ///
>                 ylabel(1 "{bf:Co-partisan candidate}" 2 "Bribes" 3 "Bribes common" ///
>                         4 "Bribes but jobs" 6 "{bf:Out-partisan candidate}" ///
>                         7 "Bribes" 8 "Bribes common" 9 "Bribes but jobs", ///
>                         labgap(6pt) noticks nogrid angle(0) labsize(small)) ytitle("") ///
>                 xscale(range(-0.6(.2)0)) xlabel(#4, glcolor(gs2)) xtitle("Effect on Pr(Voting for candidate)") ///
>                 yscale(reverse alt)  ytick(2 3 4 7 8 9, grid glcolor(gs2)) title(ANOVA-based) ///
>                 xline(0, lcolor(black) lpattern(dash) lwidth(vthin)) name(g2, replace)

. gr combine g1 g2, xcommon ycommon scheme(plotplain) rows(1)

. 
. 
. *** Figure A9: Candidate Partisanship and Co-Partisan Bias Effects for Alternative Respondent Partisanship Measures
. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. * candidate co-partisanship
. * party ID
. mat ests = J(10,3,.)

. reg vote i.corrupttreat badeconomy##arias i.copartisan i.copsource2 i.female i.country [pw=weight1500], ///
>         cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(12, 4333)       =      82.21
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1103
                                                Root MSE          =      .4277

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3533974    .013401   -26.37   0.000    -.3796702   -.3271247
         Bribes common  |  -.3631275   .0134214   -27.06   0.000    -.3894404   -.3368146
       Bribes but jobs  |   -.264988   .0145094   -18.26   0.000    -.2934338   -.2365422
                        |
             badeconomy |
      Economy worsened  |  -.0676445   .0130233    -5.19   0.000    -.0931768   -.0421122
                1.arias |  -.0528361   .0153538    -3.44   0.001    -.0829373   -.0227349
                        |
       badeconomy#arias |
    Economy worsened#1  |   .0968282    .021476     4.51   0.000     .0547242    .1389322
                        |
             copartisan |
 Co-partisan candidate  |   .0775461   .0188731     4.11   0.000     .0405451     .114547
                        |
             copsource2 |
Out-partisan newspaper  |  -.0015192   .0103112    -0.15   0.883    -.0217345    .0186961
 Co-partisan newspaper  |   .0283579   .0130783     2.17   0.030     .0027178    .0539981
                        |
                 female |
                Female  |  -.0191172   .0092638    -2.06   0.039    -.0372791   -.0009553
                        |
                country |
                 Chile  |  -.0083276   .0090087    -0.92   0.355    -.0259894    .0093341
               Uruguay  |   .0021432   .0091276     0.23   0.814    -.0157516    .0200379
                        |
                  _cons |   .5735772   .0160463    35.75   0.000     .5421182    .6050361
-----------------------------------------------------------------------------------------

. lincom 1.copartisan

 ( 1)  1.copartisan = 0

------------------------------------------------------------------------------
        vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0775461   .0188731     4.11   0.000     .0405451     .114547
------------------------------------------------------------------------------

. mat ests[3,1] = r(estimate)

. mat ests[3,2] = r(lb)

. mat ests[3,3] = r(ub)

. * left-right
. reg vote i.corrupttreat badeconomy##arias i.copartisan2 i.copsource2 i.female i.country [pw=weight1500], ///
>         cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(12, 4333)       =      80.84
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1094
                                                Root MSE          =     .42791

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3535037   .0134173   -26.35   0.000    -.3798084   -.3271989
         Bribes common  |  -.3625818   .0134479   -26.96   0.000    -.3889466    -.336217
       Bribes but jobs  |  -.2649404   .0145105   -18.26   0.000    -.2933885   -.2364924
                        |
             badeconomy |
      Economy worsened  |  -.0674867   .0130241    -5.18   0.000    -.0930206   -.0419528
                1.arias |  -.0535383   .0153874    -3.48   0.001    -.0837054   -.0233712
                        |
       badeconomy#arias |
    Economy worsened#1  |   .0962192   .0214921     4.48   0.000     .0540836    .1383548
                        |
            copartisan2 |
 Co-partisan candidate  |   .0336265   .0118851     2.83   0.005     .0103256    .0569273
                        |
             copsource2 |
Out-partisan newspaper  |   .0002461   .0103776     0.02   0.981    -.0200993    .0205915
 Co-partisan newspaper  |   .0262136   .0132277     1.98   0.048     .0002804    .0521467
                        |
                 female |
                Female  |  -.0190661   .0092779    -2.06   0.040    -.0372555   -.0008767
                        |
                country |
                 Chile  |  -.0080888   .0089955    -0.90   0.369    -.0257245     .009547
               Uruguay  |   .0083606   .0089349     0.94   0.349    -.0091563    .0258776
                        |
                  _cons |   .5696447   .0162696    35.01   0.000     .5377479    .6015414
-----------------------------------------------------------------------------------------

. mat tab = r(table)

. lincom 1.copartisan2

 ( 1)  1.copartisan2 = 0

------------------------------------------------------------------------------
        vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         (1) |   .0336265   .0118851     2.83   0.005     .0103256    .0569273
------------------------------------------------------------------------------

. mat ests[4,1] = r(estimate)

. mat ests[4,2] = r(lb)

. mat ests[4,3] = r(ub)

. * co-partisan bias
. reg vote i.corrupttreat##i.copartisan badeconomy##arias i.copsource2 i.female i.country [pw=weight1500], ///
>         cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(15, 4333)       =      66.34
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1110
                                                Root MSE          =     .42759

                                                      (Std. Err. adjusted for 4,334 clusters in uniq_id)
--------------------------------------------------------------------------------------------------------
                                       |               Robust
                                  vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
---------------------------------------+----------------------------------------------------------------
                          corrupttreat |
                               Bribes  |  -.3462265   .0139115   -24.89   0.000    -.3735001   -.3189529
                        Bribes common  |     -.3582   .0139468   -25.68   0.000    -.3855428   -.3308572
                      Bribes but jobs  |  -.2561961   .0149986   -17.08   0.000    -.2856011   -.2267911
                                       |
                            copartisan |
                Co-partisan candidate  |   .1550649   .0416301     3.72   0.000     .0734486    .2366811
                                       |
               corrupttreat#copartisan |
         Bribes#Co-partisan candidate  |  -.1085931   .0552178    -1.97   0.049    -.2168482    -.000338
  Bribes common#Co-partisan candidate  |  -.0727644   .0548904    -1.33   0.185    -.1803776    .0348489
Bribes but jobs#Co-partisan candidate  |  -.1314282   .0597686    -2.20   0.028    -.2486053   -.0142512
                                       |
                            badeconomy |
                     Economy worsened  |  -.0676367   .0130133    -5.20   0.000    -.0931494    -.042124
                               1.arias |  -.0526438   .0153534    -3.43   0.001    -.0827443   -.0225432
                                       |
                      badeconomy#arias |
                   Economy worsened#1  |   .0966118   .0214637     4.50   0.000     .0545319    .1386916
                                       |
                            copsource2 |
               Out-partisan newspaper  |  -.0006011   .0103238    -0.06   0.954    -.0208411    .0196389
                Co-partisan newspaper  |   .0289511   .0130758     2.21   0.027     .0033158    .0545863
                                       |
                                female |
                               Female  |  -.0190311   .0092637    -2.05   0.040    -.0371927   -.0008694
                                       |
                               country |
                                Chile  |  -.0084767   .0090123    -0.94   0.347    -.0261455     .009192
                              Uruguay  |   .0021014   .0091422     0.23   0.818    -.0158219    .0200247
                                       |
                                 _cons |   .5677386   .0163004    34.83   0.000     .5357816    .5996957
--------------------------------------------------------------------------------------------------------

. margins, dydx(2.corrupttreat) at(copartisan = (0 1)) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat

1._at        : copartisan      =           0

2._at        : copartisan      =           1

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |  -.3462265   .0139115   -24.89   0.000    -.3735001   -.3189529
             2  |  -.4548196   .0531942    -8.55   0.000    -.5591075   -.3505316
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. forval i = 3/4 {
  2.         local j = `i'*2
  3.         mat ests[`j',1] = tab[1,`i']
  4.         mat ests[`j',2] = tab[5,`i']
  5.         mat ests[`j',3] = tab[6,`i']
  6.         }

. reg vote i.corrupttreat##i.copartisan2 badeconomy##arias i.copsource2 i.female i.country [pw=weight1500], ///
>         cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(15, 4333)       =      64.65
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1095
                                                Root MSE          =     .42797

                                                      (Std. Err. adjusted for 4,334 clusters in uniq_id)
--------------------------------------------------------------------------------------------------------
                                       |               Robust
                                  vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
---------------------------------------+----------------------------------------------------------------
                          corrupttreat |
                               Bribes  |  -.3495742   .0150665   -23.20   0.000    -.3791122   -.3200361
                        Bribes common  |  -.3601885   .0152116   -23.68   0.000     -.390011    -.330366
                      Bribes but jobs  |  -.2658856   .0161623   -16.45   0.000     -.297572   -.2341992
                                       |
                           copartisan2 |
                Co-partisan candidate  |   .0402381   .0270703     1.49   0.137    -.0128335    .0933097
                                       |
              corrupttreat#copartisan2 |
         Bribes#Co-partisan candidate  |  -.0195175   .0341793    -0.57   0.568    -.0865263    .0474914
  Bribes common#Co-partisan candidate  |  -.0115026    .034319    -0.34   0.738    -.0787854    .0557802
Bribes but jobs#Co-partisan candidate  |   .0049743   .0373009     0.13   0.894    -.0681545    .0781031
                                       |
                            badeconomy |
                     Economy worsened  |  -.0677447   .0130226    -5.20   0.000    -.0932756   -.0422138
                               1.arias |  -.0536284   .0153902    -3.48   0.000    -.0838011   -.0234557
                                       |
                      badeconomy#arias |
                   Economy worsened#1  |   .0965547   .0214897     4.49   0.000     .0544239    .1386856
                                       |
                            copsource2 |
               Out-partisan newspaper  |   .0001626   .0103803     0.02   0.988    -.0201882    .0205134
                Co-partisan newspaper  |   .0262948   .0132259     1.99   0.047     .0003654    .0522243
                                       |
                                female |
                               Female  |  -.0189786    .009283    -2.04   0.041     -.037178   -.0007791
                                       |
                               country |
                                Chile  |  -.0081079   .0089973    -0.90   0.368    -.0257472    .0095314
                              Uruguay  |   .0082724   .0089293     0.93   0.354    -.0092337    .0257785
                                       |
                                 _cons |   .5683634   .0170526    33.33   0.000     .5349315    .6017953
--------------------------------------------------------------------------------------------------------

. margins, dydx(2.corrupttreat) at(copartisan = (0 1)) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat

1._at        : copartisan2     =           0

2._at        : copartisan2     =           1

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |  -.3495742   .0150665   -23.20   0.000    -.3791122   -.3200361
             2  |  -.3690917   .0304499   -12.12   0.000    -.4287891   -.3093943
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. forval i = 3/4 {
  2.         local j = `i'*2+1
  3.         mat ests[`j',1] = tab[1,`i']
  4.         mat ests[`j',2] = tab[5,`i']
  5.         mat ests[`j',3] = tab[6,`i']
  6.         }

. clear

. svmat ests      
number of observations will be reset to 10
Press any key to continue, or Break to abort
number of observations (_N) was 0, now 10

. forval i = 1/3 {
  2.         replace ests`i' = 0 in 2
  3.         }
(1 real change made)
(1 real change made)
(1 real change made)

. gen alt = 0

. foreach x in 4 7 9 {
  2.         replace alt = 1 in `x'
  3.         }
(1 real change made)
(1 real change made)
(1 real change made)

. gen n = .
(10 missing values generated)

. for any 1.1 1.3 1.5 1.7 2.1 2.3 2.5 2.7 2.9 3.1 \ any 1 2 3 4 5 6 7 8 9 10: replace n = X in Y

->  replace n = 1.1 in 1
(1 real change made)

->  replace n = 1.3 in 2
(1 real change made)

->  replace n = 1.5 in 3
(1 real change made)

->  replace n = 1.7 in 4
(1 real change made)

->  replace n = 2.1 in 5
(1 real change made)

->  replace n = 2.3 in 6
(1 real change made)

->  replace n = 2.5 in 7
(1 real change made)

->  replace n = 2.7 in 8
(1 real change made)

->  replace n = 2.9 in 9
(1 real change made)

->  replace n = 3.1 in 10
(1 real change made)

. gen base = 0

. replace base = 1 in 2
(1 real change made)

. twoway (rcap ests2 ests3 n if alt == 0, msize(0) lcol(black) horizontal) ///
>                 (scatter n ests1 if alt == 0, mcol(black) msize(medium) msymbol(T)) ///
>                 (rcap ests2 ests3 n if alt == 1, msize(0) lcol(gray) horizontal) ///
>                 (scatter n ests1 if alt == 1, mcol(gray) msize(medium) msymbol(S)) ///
>                 (scatter n ests1 if base == 1, mcol(black) msize(medlarge) msymbol(O)), ///
>                 legend(ring(0) pos(5) order(2 "Party ID" 4 "Left-right")) scheme(plotplain) ///
>                 ylabel(1.1 "{bf: Candidate partisandship}" 1.3 "Out-partisan candidate ({it:reference})" ///
>                                 1.6 "Co-partisan candidate" ///
>                                 2.1 "{bf: Co-partisan bias}" 2.4 "Bribes & out-partisan candidate" ///
>                                 2.8 "Bribes & co-partisan candidate", ///
>                                 labgap(6pt) noticks nogrid angle(0) labsize(small)) ytitle("") ///
>                 xscale(range(-0.6(.1)0)) xlabel(#4, glcolor(gs2)) xtitle("Effect on Pr(Voting for candidate)") ///
>                 yscale(reverse )  ytick(1.3 1.5 1.7 2.3 2.5 2.7 2.9, grid glcolor(gs2) notick) ///
>                 xline(0, lcolor(black) lpattern(dash) lwidth(vthin))

.                 
. 
. *** Figure A10: Left-Party vs. Right-Party Candidate Effects
. use analysis-data, clear
(�AmericasBarometer, LAPOP; created 22 Sep 2017; type: notes list)

. gen lr = 0

. replace lr = 1 if leftid==1 & party==1
(400 real changes made)

. replace lr = 2 if rightid== 1 & party==2
(209 real changes made)

. mat ests = J(9,3,.)

. * candidate partisanship
. reg vote i.corrupttreat badeconomy##arias i.lr i.copsource2 i.female i.country [pw=weight1500], ///
>         cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(13, 4333)       =      76.12
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1103
                                                Root MSE          =     .42771

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3534196   .0133998   -26.37   0.000    -.3796901   -.3271491
         Bribes common  |  -.3630573   .0134255   -27.04   0.000    -.3893781   -.3367364
       Bribes but jobs  |  -.2648837   .0145126   -18.25   0.000    -.2933359   -.2364315
                        |
             badeconomy |
      Economy worsened  |  -.0677506   .0130212    -5.20   0.000    -.0932787   -.0422224
                1.arias |  -.0528917   .0153543    -3.44   0.001     -.082994   -.0227894
                        |
       badeconomy#arias |
    Economy worsened#1  |   .0970083   .0214712     4.52   0.000     .0549137    .1391029
                        |
                     lr |
                     1  |   .0697368   .0226304     3.08   0.002     .0253696     .114104
                     2  |    .092994   .0311129     2.99   0.003     .0319967    .1539912
                        |
             copsource2 |
Out-partisan newspaper  |  -.0014396   .0103127    -0.14   0.889    -.0216577    .0187785
 Co-partisan newspaper  |   .0283901   .0130779     2.17   0.030     .0027507    .0540295
                        |
                 female |
                Female  |   -.019124   .0092654    -2.06   0.039     -.037289    -.000959
                        |
                country |
                 Chile  |  -.0081455   .0090137    -0.90   0.366     -.025817     .009526
               Uruguay  |   .0022659   .0091311     0.25   0.804    -.0156357    .0201675
                        |
                  _cons |   .5734332   .0160528    35.72   0.000     .5419614     .604905
-----------------------------------------------------------------------------------------

. margins, dydx(lr) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 1.lr 2.lr

------------------------------------------------------------------------------
             |            Delta-method
             |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          lr |
          1  |   .0697368   .0226304     3.08   0.002     .0253696     .114104
          2  |    .092994   .0311129     2.99   0.003     .0319967    .1539912
------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. forval i = 2/3 {
  2.         local j = `i'+1
  3.         mat ests[`j',1] = tab[1,`i']
  4.         mat ests[`j',2] = tab[5,`i']
  5.         mat ests[`j',3] = tab[6,`i']
  6.         }

. * co-partisan bias
. reg vote i.corrupttreat##i.lr badeconomy##arias i.copsource2 i.female i.country [pw=weight1500], ///
>         cl(uniq_id)
(sum of wgt is 8,355.07181036472)

Linear regression                               Number of obs     =      8,668
                                                F(19, 4333)       =      52.67
                                                Prob > F          =     0.0000
                                                R-squared         =     0.1112
                                                Root MSE          =     .42766

                                       (Std. Err. adjusted for 4,334 clusters in uniq_id)
-----------------------------------------------------------------------------------------
                        |               Robust
                   vote |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
           corrupttreat |
                Bribes  |  -.3462263   .0139146   -24.88   0.000     -.373506   -.3189466
         Bribes common  |  -.3582019   .0139502   -25.68   0.000    -.3855514   -.3308525
       Bribes but jobs  |  -.2561959   .0150019   -17.08   0.000    -.2856073   -.2267845
                        |
                     lr |
                     1  |   .1304615   .0524021     2.49   0.013     .0277265    .2331965
                     2  |   .1989498    .062572     3.18   0.001     .0762766     .321623
                        |
        corrupttreat#lr |
              Bribes#1  |  -.0997511   .0680074    -1.47   0.143    -.2330805    .0335782
              Bribes#2  |  -.1263209   .0892336    -1.42   0.157    -.3012644    .0486226
       Bribes common#1  |  -.0372875   .0685508    -0.54   0.587    -.1716821    .0971071
       Bribes common#2  |   -.139861   .0863169    -1.62   0.105    -.3090863    .0293643
     Bribes but jobs#1  |  -.1081993   .0726031    -1.49   0.136    -.2505386      .03414
     Bribes but jobs#2  |  -.1717665   .1017849    -1.69   0.092     -.371317     .027784
                        |
             badeconomy |
      Economy worsened  |  -.0676478   .0130209    -5.20   0.000    -.0931754   -.0421202
                1.arias |  -.0527225   .0153552    -3.43   0.001    -.0828265   -.0226185
                        |
       badeconomy#arias |
    Economy worsened#1  |   .0965561   .0214662     4.50   0.000     .0544714    .1386408
                        |
             copsource2 |
Out-partisan newspaper  |  -.0005824   .0103262    -0.06   0.955     -.020827    .0196623
 Co-partisan newspaper  |   .0288013   .0130774     2.20   0.028     .0031629    .0544398
                        |
                 female |
                Female  |  -.0190871   .0092753    -2.06   0.040    -.0372713   -.0009028
                        |
                country |
                 Chile  |  -.0084271   .0090233    -0.93   0.350    -.0261175    .0092633
               Uruguay  |   .0019951   .0091452     0.22   0.827    -.0159343    .0199244
                        |
                  _cons |   .5678619   .0163182    34.80   0.000     .5358699     .599854
-----------------------------------------------------------------------------------------

. margins, dydx(2.corrupttreat) at(lr = (0 1 2)) post

Average marginal effects                        Number of obs     =      8,668
Model VCE    : Robust

Expression   : Linear prediction, predict()
dy/dx w.r.t. : 2.corrupttreat

1._at        : lr              =           0

2._at        : lr              =           1

3._at        : lr              =           2

---------------------------------------------------------------------------------
                |            Delta-method
                |      dy/dx   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.corrupttreat  |  (base outcome)
----------------+----------------------------------------------------------------
2.corrupttreat  |
            _at |
             1  |  -.3462263   .0139146   -24.88   0.000     -.373506   -.3189466
             2  |  -.4459774   .0664571    -6.71   0.000    -.5762674   -.3156875
             3  |  -.4725472   .0878875    -5.38   0.000    -.6448516   -.3002428
---------------------------------------------------------------------------------
Note: dy/dx for factor levels is the discrete change from the base level.

. mat tab = r(table)

. forval i = 4/6 {
  2.         local j = `i' + 2
  3.         mat ests[`j',1] = tab[1,`i']
  4.         mat ests[`j',2] = tab[5,`i']
  5.         mat ests[`j',3] = tab[6,`i']
  6.         }

. clear

. svmat ests      
number of observations will be reset to 9
Press any key to continue, or Break to abort
number of observations (_N) was 0, now 9

. forval i = 1/3 {
  2.         replace ests`i' = 0 in 2
  3.         }
(1 real change made)
(1 real change made)
(1 real change made)

. gen alt = 0

. foreach x in 4 8 {
  2.         replace alt = 1 in `x'
  3.         }
(1 real change made)
(1 real change made)

. foreach x in 2 6 {
  2.         replace alt = 2 in `x'
  3.         }
(1 real change made)
(1 real change made)

. gen n = .
(9 missing values generated)

. for any 1.1 1.3 1.5 1.7 2.1 2.3 2.5 2.7 2.9 \ any 1 2 3 4 5 6 7 8 9: replace n = X in Y

->  replace n = 1.1 in 1
(1 real change made)

->  replace n = 1.3 in 2
(1 real change made)

->  replace n = 1.5 in 3
(1 real change made)

->  replace n = 1.7 in 4
(1 real change made)

->  replace n = 2.1 in 5
(1 real change made)

->  replace n = 2.3 in 6
(1 real change made)

->  replace n = 2.5 in 7
(1 real change made)

->  replace n = 2.7 in 8
(1 real change made)

->  replace n = 2.9 in 9
(1 real change made)

. twoway (rcap ests2 ests3 n if alt == 0, msize(0) lcol(black) horizontal) ///
>                 (scatter n ests1 if alt == 0, mcol(black) msize(medium) msymbol(T)) ///
>                 (rcap ests2 ests3 n if alt == 1, msize(0) lcol(black) horizontal) ///
>                 (scatter n ests1 if alt == 1, mcol(black) msize(medium) msymbol(S)) ///
>                 (rcap ests2 ests3 n if alt == 2, msize(0) lcol(gray) horizontal) ///
>                 (scatter n ests1 if alt == 2, mcol(gray) msize(medium) msymbol(O)), ///
>                 legend(ring(0) pos(5) order(2 "Left party" 4 "Right party" 6 "Independent")) scheme(plotplain) ///
>                 ylabel(1.1 "{bf: Candidate partisandship}" 1.3 "Independent ({it:reference})" ///
>                                 1.5 "Left party co-partisan" 1.7 "Right party co-partisan" ///
>                                 2.1 "{bf: Co-partisan bias}" 2.3 "Bribes & independent candidate" ///
>                                 2.5 "Bribes & left party co-partisan candidate" ///
>                                 2.7 "Bribes & right party co-partisan candidate", ///
>                                 labgap(6pt) noticks nogrid angle(0) labsize(small)) ytitle("") ///
>                 xscale(range(-0.6(.1)0)) xlabel(#4, glcolor(gs2)) xtitle("Effect on Pr(Voting for candidate)") ///
>                 yscale(reverse )  ytick(1.3 1.5 1.7 2.3 2.5 2.7, grid glcolor(gs2)) ///
>                 xline(0, lcolor(black) lpattern(dash) lwidth(vthin))

. 
. gr drop _all

. 
. log close
      name:  <unnamed>
       log:  C:\Users\marko\Dropbox\07_LatinAmerica_Corruption_Paper\LAPOP\replication\appendix-log.log
  log type:  text
 closed on:  12 May 2020, 09:31:39
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
